5GNOW D3.3 Final

The 5GNOW Project Consortium groups the following organizations:
Partner Name Short name Country

FRAUNHOFER-GESELLSCHAFT ZUR FOERDERUNG DER ANGEWANDTEN
FORSCHUNG E.V. HHI Germany
ALCATEL LUCENT DEUTSCHLAND AG ALUD Germany
COMMISSARIAT A L ENERGIE ATOMIQUE ET AUX ENERGIES
ALTERNATIVES CEA France
IS-WIRELESS ISW Poland
NATIONAL INSTRUMENTS NI Hungary
TECHNISCHE UNIVERSITT DRESDEN TUD Germany
Final 5GNOW Transceiver and frame structure concept

Abstract:
The screening process of T3.1 will generate
D3.3a candidate list of waveforms and signal formats,
described in this IR3.1.
5GNOW_D3.3_v1.0.docx
Version: 1.0
Last Update: 4/5/2015
Distribution Level: CO
Distribution level
PU = Public,
RE = Restricted to a group of the specified Consortium,
PP = Restricted to other program participants (including Commission Services),
CO= Confidential, only for members of the 5GNOW Consortium (including the Commission Services)
D3.3
The 5GNOW Project Consortium groups the following organizations:
Partner Name Short name Country

FRAUNHOFER-GESELLSCHAFT ZUR FOERDERUNG DER ANGEWANDTEN
FORSCHUNG E.V. HHI Germany
ALCATEL LUCENT DEUTSCHLAND AG ALUD Germany
COMMISSARIAT LENERGIE ATOMIQUE ET AUX ENERGIES
ALTERNATIVES CEA France
IS-WIRELESS ISW Poland
NATIONAL INSTRUMENTS NI Hungary
TECHNISCHE UNIVERSITT DRESDEN TUD Germany
Abstract:
Built upon D3.2, a final 5NOW transceiver and frame structure concept is described. Final
performance results, e.g. based on simulations, are presented. Optimal and/or reasonable
parameters are presented for the waveform types selected in D3.1 under the scenario conditions
given by WP2.
Partner Author name

HHI Gerhard Wunder, Martin Kasparick, Peter Jung
ALUD Yejian Chen, Frank Schaich, Thorsten Wild
CEA Vincent Berg, Nicolas Cassiau, Jean-Baptiste Dor, Dimitri Ktnas
ISW ISW
NI NI
TUD Ivan Gaspar, Nicola Michailow
5GNOW Page: 2 of 95
FP7 ICT GA 318555
D3.3
The research leading to these results has received funding from the European Community's Seventh
Framework Programme (FP7/2007-2013) under grant agreement n 318555
5GNOW Page: 3 of 95
FP7 ICT GA 318555
D3.3
Document Identity
Title: Final 5GNOW Transceiver and frame structure concept

WP: WP3 Physical layer design
WP Leader ALUD
Main Editor Thorsten Wild, Gerhard Wunder
Number: D3.3
File name: 5GNOW_D3.3_v1.0.docx
Last Update: Monday, May 04, 2015
Revision History
No. Version Edition Author(s) Date

1 0.1 Thorsten Wild (ALUD) 10.10.15
Comments: Providing initial template and skeleton
2 0.2 Thorsten Wild (ALUD) 16.02.15
Comments: Merged different partner inputs
3 0 Gerhard Wunder (HHI) 31.03.15
Comments: Final review
4 0
Comments:
5 0
Comments:
6 1
Comments:
7
Comments:
8
Comments:
9
Comments:
10
Comments:
11
Comments:
12
Comments:
13
Comments:
14
Comments:
15
Comments:
5GNOW Page: 4 of 95
FP7 ICT GA 318555
D3.3
Table of Contents
1 INTRODUCTION ....................................................................................................................................... 8
1.1 SPORADIC TRAFFIC..................................................................................................................................... 8
1.2 SPECTRAL AND TEMPORAL FRAGMENTATION................................................................................................... 9
1.3 REAL-TIME CONSTRAINTS ............................................................................................................................ 9
1.4 NEW WAVEFORMS AND MULTIPLE ACCESS THE GOLDEN AIR INTERFACE ....................................................... 10
2 5GNOW WAVEFORMS........................................................................................................................... 12
2.1 WAVEFORM CANDIDATES ......................................................................................................................... 12
2.2 GFDM ................................................................................................................................................. 12
2.2.1 System Description and Properties ................................................................................................ 12
2.2.2 Fast computation of the receive filter ........................................................................................... 15
2.2.3 Multi-User Time-Reversal Space-Time Coding for GFDM .............................................................. 16
2.3 UFMC ................................................................................................................................................. 18
2.3.1 UFMC Frequency- and Time Domain Properties ........................................................................... 20
2.3.2 Channel estimation and equalization ............................................................................................ 21
2.3.3 Native MIMO/CoMP support ........................................................................................................ 21
2.3.4 Support for multi-user sounding and multiplex of small control resource element groups .......... 22
2.3.5 Adaptivity Potential ....................................................................................................................... 22
2.4 FBMC .................................................................................................................................................. 23
2.4.1 FBMC principles ............................................................................................................................. 23
2.4.2 FBMC receiver ............................................................................................................................... 26
2.5 BFDM.................................................................................................................................................. 27
2.5.1 Transmitter and receiver Structure ............................................................................................... 27
2.5.2 Pulse design ................................................................................................................................... 29
2.5.3 Synchronization and Equalization ................................................................................................. 33
3 5GNOW FRAME STRUCTURE AND MULTIPLE ACCESS ............................................................................ 35
3.1 UNIFIED FRAME STRUCTURE ..................................................................................................................... 35
3.2 LAYERING .............................................................................................................................................. 36
3.2.1 A Brief Review of IDMA ................................................................................................................. 36
3.2.2 Involving IDMA in UFMC ............................................................................................................... 38
3.3 RELAXED SYNCHRONIZATION APPROACHES AND AUTONOMOUS TIMING ADVANCE .............................................. 39
3.3.1 Relaxed Synchronization with UFMC ............................................................................................. 39
3.3.2 Relaxed Synchronization with FBMC ............................................................................................. 40
3.4 D-PRACH............................................................................................................................................. 41
3.4.1 5GNOW one shot transmission concept .................................................................................... 41
4 PERFORMANCE RESULTS ....................................................................................................................... 44
4.1 WAVEFORM PERFORMANCE AND PARAMETER OPTIMIZATION ......................................................................... 44
4.1.1 Generalized frequency division multiplexing ................................................................................. 44
4.1.2 UFMC ............................................................................................................................................. 52
4.1.3 FBMC ............................................................................................................................................. 57
4.2 FRAME STRUCTURE AND MULTIPLE ACCESS PERFORMANCE ............................................................................ 67
4.2.1 IDMA and Superimposed Pilots ..................................................................................................... 67
4.2.1.1 Signal Model ............................................................................................................................. 68
4.2.1.2 Detection and Estimation ......................................................................................................... 70
4.2.1.3 Iterative Processing .................................................................................................................. 71
4.2.1.4 Simulation and Numerical Results ............................................................................................ 72
4.2.1.5 Summary................................................................................................................................... 75
4.2.2 D-PRACH Performance .................................................................................................................. 76
5 COMPARISON AND DISCUSSION OF SELECTED WAVEFORMS ................................................................ 84
5GNOW Page: 5 of 95
FP7 ICT GA 318555
D3.3
6 CONCLUSION......................................................................................................................................... 89
7 ABBREVIATIONS AND REFERENCES ....................................................................................................... 90
5GNOW Page: 6 of 95
FP7 ICT GA 318555
D3.3
Executive Summary
Starting from the main drivers of 5G, sporadic traffic, spectral and temporal fragmentation and
real-time constraints with the vision of supporting a single unified air interface, D3.3 presents the
building blocks of the final 5GNOW transceiver and frame structure concept. General waveform
considerations are discussed in the light of the Gabor theory. Then the four 5GNOW waveform
candidate technologies GFDM, UFMC, FBMC and BFDM are described in detail, summarizing the
available results.These waveforms are supporting and enabling the Unified Frame Sturcture. The
Unified Frame Structure concept is the heart of the 5GNOW frame design, designed for supporting
various heterogeneous traffic and device types in parallel.Multiple superimposed signal layers are
supported by advanced multiple access techniques like IDMA, which can be combined with 5GNOW
waveforms in order to support efficient multiuser detection mechanisms.Techniques and performance
results are presented for relaxed synchronization support. Directly connect to the Unified Frame
structure is the introduction of one-shot-transmission using a new proposed physical random access
channel for data transmission (D-PRACH). Performance results for the candidate transceiver
approaches are provided, including optimized and/or reasonable waveform parameters.The
waveforms are discussed and compared with each other.The conclusion is that 5GNOW has provided
a powerful waveform and frame structure toolbox for the 5GPPP research work towards the
Horizon 2020, coming along with optimized/reasonable parameter settings and performance results.
The further system design steps taken in future 5GPP projects can built upon this vast number of
available 5GNOW technologies and results for waveforms and frame structure of a new 5G air
interface, paving the way for 5G standardization.
5GNOW Page: 7 of 95
FP7 ICT GA 318555
D3.3
1 Introduction
The main hypothesis of 5GNOW is that the underlying design principles synchronism and
orthogonality of the PHY layer of todays LTE-A radio access network constitute a major obstacle for
the envisioned service architecture.
Synchronism means that the senders operate with a common clock for their processing.
Orthogonality means that no crosstalk occurs in the receivers waveform detection process. Often,
both are related such that some rough synchronization is required. LTE-A OFDM modulation keeps
the subcarrier waveforms orthogonal even after the channel, provided the DFT window can be
properly adjusted by suitable synchronization mechanism, which is then near optimal processing in a
single cell provided that a capacity achieving scheme such as superposition coding is used per
subcarrier. However, as soon as the orthogonality is destroyed, e.g., due to random channel access
or multi-cell operation, the distortion accumulates without bounds in OFDM. This is due to the so-
called reproducing Dirichlet kernel sin(Nx)/sin(x) of OFDM which quickly approaches the sin(x)/x
kernel for large N where N is the number of subcarriers. For such kernel, it is well-known, that the
amplification of small errors e.g., due to sampling or frequency offsets, is not independent of N and
can grow with order log(N). Hence, we believe it is better to abandon strict orthogonality partially or
altogether and control the impairments instead. Let us discuss several examples in the following
[WJK+14].
1.1 Sporadic traffic
Sporadic traffic generating devices such as machine-type communications (MTC) in the Internet of
things (IoT) should not be forced to be integrated into the bulky synchronization procedure of LTE-A
PHY layer random access. Instead, ideally, they awake occasionally, and then they should transmit
their messages right away and only coarsely synchronized. By doing so MTC traffic would be removed
from standard uplink data pipes allowing for drastically reduced signalling overhead. Therefore,
alleviating the synchronism requirements can significantly improve operational capabilities and
network performance as well as user experience and lifetime of autonomous MTC nodes.
Interestingly, sporadic access poses another significant challenge to mobile access networks due to
an operation known as fast dormancy. Fast dormancy is used by smartphone manufacturers to save
battery power by using the feature that a mobile can break ties to the network individually and as
soon as a data piece is delivered the smartphone changes from active into idle state. Consequently,
when the mobile has to deliver more pieces of data it will always go through the complete
synchronization procedure again. Actually, this can happen several hundred times a day resulting in
significant control signalling growth and network congestion threat. A rough estimation yields that 2k
control resource elements (i.e. a subcarrier) are necessary to deliver one data resource element. We
conclude that sporadic traffic needs to be carried by new, possibly non-orthogonal waveforms for
asynchronous signalling in the uplink and specifically in an uplink random access channel (RACH).
In this deliverable we propose so-called one shot transmission together with Bi-orthogonal
Frequency Division Multiplexing to meet 5G scalability requirements.
5GNOW Page: 8 of 95
FP7 ICT GA 318555
D3.3
1.2 Spectral and temporal fragmentation
Due to its fragmentation, spectrum is scarce and expensive but also underutilized: this is commonly
referred to as the spectrum paradox. Therefore, carrier aggregation will be implemented to achieve
much higher rates by variably aggregating non-contiguous frequency bands [NGV11]. Carrier
aggregation implies the use of separate RF front ends accessing different channels thereby
reinforcing the attraction of isolated frequency bands such as the L-Band. Actually, the search for
new spectrum is very active in Europe and in the USA in order to provide mobile broadband
expansion. It includes the opportunistic use of spectrum, which has been an interesting research area
in wireless communications in the past decade. Moreover, techniques to detect and assess channel
vacancy using cognitive radio could well make new business models possible in the future. The first
real implementation will start with the exploration of TV white spaces in the USA. Combined with the
preparation of the on-going regulatory framework in Europe, opportunistic use of spectrum and
spectrum agility can address a 5G market if it efficiently implements the requirements for protection
of legacy systems such as low out-of-band radiation [NGV11].
LTE-A imposes generous guard bands to other legacy networks to satisfy spectral mask requirements
which either severely deteriorate spectral efficiency or even prevent band usage at all, which is again
an artefact of the strict orthogonality and synchronism constraints of OFDM. Moreover, in a scenario
with uncoordinated interference from Pico- or Femto-cells and highly overlapping coverage, it seems
illusive to provide the degree of coordination to maintain synchronism and orthogonality in the
network calling for new waveforms as well. In addition to spectral fragmentation, temporal
fragmentation is another key issue, e.g., due to one shot sporadic access in the asynchronous
uplink RACH. Notably, asynchronous signaling matters also in the downlink in the context of
cooperative multipoint (CoMP).
In conclusion, such 5G scenarios where multiple users are allocated a pool of frequencies with
relaxed (or even no) synchronization in time must be addressed by new waveforms. Such waveforms
must implement sharp frequency notches and tight spectral masks in order not to interfere with
other legacy systems, must be robust to asynchronous signalling and handle un-coordinated
interference. Traditional OFDM schemes are not suited due to the inflexible handling of guard
intervals (GIs) cyclic prefixes (CPs) or cyclic suffixes (CS) as well as poor spectral localization.
In this deliverable we discuss waveforms achieving 100x better localization (e.g., 35 dB side lobe with
LTE-A OFDM compared to 55dB side lobe with Filter Bank Multi-Carrier (FBMC) [FB11]) which makes
then a real difference in fragmented spectrum and CoMP scenarios.
1.3 Real-time constraints
4G systems offer latencies of multiple 10ms between terminal and base station which originate from
resource scheduling, frame processing, re-transmission procedures, etc. However, future application
scenarios such as the Tactile Internet scenario require ultra-low latency matched with the human
tactile sense. In such an environment, a massive number of distributed sensors and actuators will be
connected to enable real-time tactile interaction in an augmented way. Sharing the medium
becomes an additional challenge and imposes short wake up cycles on the nodes and the use of
burst transmission. Instead of consuming spectrum and power resources by introducing
sophisticated algorithms to reach synchronism, an asynchronous approach appears promising.
5GNOW Page: 9 of 95
FP7 ICT GA 318555
D3.3
In order to achieve ultra-low latency, each and every element of the communication and control
chain must be optimized. Focusing on the PHY layer, LTE-A system supports different granularity of
scheduling resources in a fixed transmission time interval (TTI) of 1ms. TTI represents an inherent
lower bound of the LTE-A systems PHY latency. Clearly, as the time budget on PHY layer in the
Tactile Internet scenario is 100s maximum, frame duration must be reduced and LTE-A with its
OFDM symbol duration of 67s is not an option. In order to discuss possible alternatives assume
20s symbol duration. Considering e.g., a 1km cell range, the expected delay spread is around 3s,
and, thus, 4s CP is required to ensure an inter-symbol interference (ISI) free scenario. Hence, use of
conventional OFDM entails 20% loss in spectral efficiency. On the other hand, a non-orthogonal
waveform which allows for transmitting multiples symbols with a single CP relaxes such strict time
domain requirements.
Another major drawback caused by short frames is the fixed bandwidth increment required to keep a
given throughput. A flexible non-orthogonal multicarrier waveform allowing also for inter-carrier
interference (ICI) can use flexible subcarrier spacing to accommodate the necessary bandwidth.
Alternatively, non-contiguous spectrum can be aggregated again enabled by the low out-of-band
emissions of the non-orthogonal waveform.
Short frames have also positive impact on mobility support or operational frequencies. LTE-A has
been designed to support Doppler spread of 100Hz caused by 50km/h mobility for the respective
LTE-A carrier frequency. By reducing the frame duration it is possible to support either higher
mobility or to operate in a higher frequencies range. Finally, a short frame brings benefits to upper
protocol layers: Although the low latency requirements of real-time applications demands for a
robust PHY layer to avoid retransmissions of the frame, applications may desire acknowledged
signaling. A short frame will enable the implementation of less time-consuming retransmissions
algorithms.
Summarizing, although OFDM could be tuned to address different granularity of scheduling

resources, there is no mode in the current LTE-A standard that can adapt to the latency requirements
of real-time services running on top. If the symbol duration is reduced to achieve very short
roundtrip delays, the GIs cannot be scaled accordingly without severely compromising spectral
efficiency or cell size. We conclude that required flexibility can only be achieved with new
waveforms.
In this deliverable we discuss Generalized Frequency Division Multiplexing (GFDM) to achieve such
flexibility in the frame design.
1.4 New waveforms and multiple access the Golden Air Interface
We have discussed that the 5G services will be very much different with different requirements. One
alternative for fulfilling 5G targets is to introduce separate specialized air interfaces on dedicated
bands. The potential drawbacks are that the spectrum will be inefficiently used (e.g. due to lack of
multiplexing gain) and multiple parallel implementations need to be supported and maintained at
network elements and devices (increasing cost).
5GNOW Page: 10 of 95
FP7 ICT GA 318555
D3.3
A second option is to aim for a Golden Air Interface [SSS14]. This means a single air interface with
modular design, which is adaptable and reconfigurable and can be used efficiently for the different
services.
The role of the waveform in this Golden Air Interface is to support the flexibility and modular design.
5GNOW waveforms, due to better spectral properties than OFDM, thus have the potential to enable
the support of very heterogeneous requirements on parallel subbands, as they provide a better
spectral separation of heterogeneous multi-carrier parameter sets and different accuracy levels of
time-frequency synchronization, which would cause inter-carrier interference in OFDM.
In this deliverable we discuss Universal Filtered Multicarrier (UFMC) which is a promising technique
for such requirements.
FP7 ICT GA 318555
D3.3
2 5GNOW Waveforms
2.1 Waveform candidates
In 5GNOW the following waveform approaches have been developed:

GFDM can be seen as a more generic block oriented filtered multicarrier system that follows
the Gabor principles. Basically, the parameterization of the waveform directly influences i)
transmitter window; ii) time-frequency grid structure; as well as iii) transform length and can
hence provide means to emulate a multitude of conventional multi-carrier systems.
In UFMC a pulse shaping filter is applied to a group of conventional OFDM subcarriers. This
approach can be also represented in the context of the Gabor frame.
FBMC-OQAM belongs to the family of filterbank based waveforms. The principles revolve
around filtering the subcarriers in the system while retaining orthogonality. As the name
suggests, the essence of this candidate waveform is offset modulation, which allows avoiding
interference between real and imaginary signal components.
BFDM directly relates to the theory of Gabor frames. Signal generation can be considered a
Gabor expansion while the bi-orthogonal receive filter constitutes a Gabor transform.
These waveforms have been thoroughly investigated within 5GNOW, each particularly related to
certain scenarios as described in detail in the next section.
2.2 GFDM
Generalized Frequency Division Multiplexing (GFDM) is an innovative multicarrier modulation

scheme where the subcarriers are individually pulse-shaped in a block structure of M subsymbols and
K subcarriers. Because all filter impulse responses are derived from a prototype baseband filter
through time and frequency circular shifts, the GFDM signal is confined to MK samples. A single cyclic
prefix (CP) for M subsymbols allows for simple frequency domain equalization. Although a new
waveform might be necessary to deal with the new requirements of the fifth generation of mobile
communication (5G), a completely disruptive approach is not interesting for operators. GFDM is a
very flexible waveform, which covers Orthogonal Frequency Division Multiplexing (OFDM) and Single
Carrier Frequency Division Multiplexing (SC-FDM) as corner cases, providing support for the two
modulation schemes used in LTE.
2.2.1 System Description and Properties
Figure 2.2.1: Block diagram of the transceiver
FP7 ICT GA 318555
D3.3
Consider the block diagram depicted in Figure 2.2.1. A data source provides the binary data vector b ,
which is encoded to obtain b c . A mapper, e.g. quadrature amplitude modulation (QAM), maps
sequences of encoded bits to symbols d of a 2-valued complex constellation. The resulting
vector d denotes a GFDM data block that contains N elements, which can be decomposed
into K groups of M symbols according
to d d 0 T , , d K 1 T
T
and d k d k , 0 , , d k , M 1 with N = KM.
T
Therein, the individual
elements d k , m correspond to the data transmitted on the kth subcarrier and in the mth subsymbol of
the block. In the GFDM modulator, each d k , m is transmitted with the corresponding pulse shape
n
j 2 k
g k , m [ n ] g n m K m o d N e K (2.2.1.1)
with n denoting the sampling index. Each g k , m [ n ] is a time and frequency shifted version of a
prototype filter g[n] , where the modulo operation makes g k , m [ n ] a circularly shifted version
of g k , 0 [ n ] and the complex exponential performs the shifting operation in frequency. The transmit
samples x x [ n ] are obtained by superposition of all transmit symbols
T
K 1 M 1
(2.2.1.2)
x[ n ] g k ,m [ n ]d k ,m , n 0, , N 1.
k 0 m0
Collecting the filter samples in a vector g k , m g k , m [ n ] allows to formulate (2.2.1.2) as

T
x Ad , (2.2.1.3)
where A is a KM KM transmitter matrix [MKL+12] with a structure according to

A g 0 ,0
g K 1,0 g 0 ,1 g K 1 ,1 g 0,M 1
g K 1, M 1 . (2.2.1.4)
Figure 2.2.2 shows three columns of an example transmitter matrix. As one can see,
g 1 , 0 [ A ] n , 2 and g 0 ,1 [ A ] n , K 1 are circularly frequency and time shifted versions of g 0 , 0 [ A ] n ,1 .
At this point, x contains the transmit samples that correspond to the GFDM data block d . Lastly, on
the transmitter side a cyclic prefix of N CP samples is added to produce x .
Transmission through a wireless channel is modelled by y H x w , where y is the received

counterpart of x . Here, H is a convolution matrix with band-diagonal structure based on a channel
impulse response h which is a realization of a Rayleigh multipath fading channel. Lastly,
w ~ N 0, 2
w
IN N
CP
denotes additive white Gaussian noise. At the receiver, time and frequency
synchronization is performed, yielding y s . Then the cyclic prefix is removed. Under the assumption
of perfect synchronization, i.e. y s y , the cyclic prefix can be utilized to simplify the model of the
wireless channel to
y Hx w (2.2.1.5)
by replacing the matrix H with the corresponding circular convolution matrix H . This allows
employing zero-forcing channel equalization as efficiently used in OFDM [Bin90]. The overall
FP7 ICT GA 318555
D3.3
transceiver equation can be written as y H A d w . Introducing z H 1 H A d H 1 w A d w

as the received signal after channel equalization, linear demodulation of the signal can be expressed
as
d B z , (2.2.1.6)
where B is a KM KM receiver matrix.
Several standard receiver options for the GFDM demodulator are readily available in literature:
The matched filter (MF) receiver B M F A maximizes the signal-to-noise ratio (SNR) per subcarrier,
H
but with the effect of introducing self-interference when a non-orthogonal transmit pulse is applied,
i.e. the scalar product g 0 , 0 , g k , m N
0 ,k 0 ,m with Kronecker delta i,j.
1
The zero-forcing (ZF) receiver B Z F A on the contrary completely removes any self-interference
at the cost of enhancing the noise. Also, there are cases in which A is ill-conditioned and thus the
inverse does not exist. The linear minimum mean square error (MMSE) receiver
1
B M M S E ( w I A A ) A
2 H H
makes a trade-off between self-interference and noise enhancement.
(a) A transmitter matrix (b) Three columns in detail
Figure 2.2.2: Illustration of the GFDM transmitter matrix for N = 28, K = 4, M = 7, using a raised cosine (RC) filter
with a=0.4.
Finally, the received symbols d are demapped to produce a sequence of bits bc at the receiver,
which are then passed to a decoder to obtain b .
From the description of the transmitter and receiver, it is clear that GFDM falls into the category of
filtered multicarrier systems. The name derives from the fact that the scheme offers more degrees of
freedom than traditional OFDM or single carrier with frequency domain equalization (SC-FDE). GFDM
turns into OFDM when M 1 , A F N and B F N , where F N is a N N Fourier matrix. SC-FDE
H
is obtained when K 1 and SC-FDM - a frequency division multiplexing of several SC-FDE signals - is
obtained when g is a Dirichlet pulse [MF13]. However, the important property that distinguishes the
proposed scheme from OFDM and SC-FDE is that, like SC-FDM, it allows dividing a given time-
frequency resource into K subcarriers and M subsymbols as depicted in Figure 2.2.3. Therefore, it is
FP7 ICT GA 318555
D3.3
possible to engineer the spectrum according to given requirements and enables pulse shaping on a
per subcarrier basis.
As a consequence, without changing the sampling rate, GFDM can be configured to cover a portion
of bandwidth either with a large number of narrow band subcarriers like in OFDM or with a small
number of subcarriers of large individual bandwidth like in SC-FDM. Further it is important to note
that although filters are introduced, GFDM is still a block based approach. These aspects are relevant
for the scheduling of users in a multiple access scenario [WJK+14] and also when targeting low
latency transmissions [Fet14]. Within each block, the signal is designed such that it exhibits a circular
structure in time and frequency domain. In combination with a cyclic prefix at the beginning of each
GFDM block, this property helps to keep transmitter and receiver complexity low [GMC+13] and
eases synchronization and equalization.
(a) OFDM (b) SC-FDE (c) SC-FDM and GFDM
Figure 2.2.3: Partitioning of time and frequency, where data occupies different resources depending on the
chosen scheme. (a) with K = N subcarriers and M = 1 subsymbols, (b) with K = 1 subcarriers
and M = N subsymbols and (c) with K = 4, M = 3 and N = 12
2.2.2 Fast computation of the receive filter

If g[n] is identified as a discrete Gabor prototype window where the data symbols dk,m are the Gabor
expansion coefficients, then, clearly, the GFDM signal is a critically sampled Gabor expansion so that
[n] is the corresponding dual window to g[n] (with the same time-frequency cyclic shift structure)
and the received filters represent the Gabor transform of the equalized received sequence. One
important practical application of this theory is the computation of the prototype receiver filter.
Since [n] is the dual to g[n], the prototype receiver filter can be obtained by
1
[, ] = (( 1 )(,) ) ( (,)
), (2.2.2.1)
( ) [, ]
where
1

( (,) )[, ] = ( + ) 2 (2.2.2.2)

=0
is the Discrete Zak Transform (DZT) of a periodic sequence g[n] and ( 1 )(,) is the inverse DZT.
From (2.2.2.1) and (2.2.2.2), it is clear that the evaluation of the receiver filters does not require a
large computation based on all possible waveforms, but the computationally efficient DZT transform
pair can provide the prototype receiver filter only based on the transmit prototype filter.
The consequence of pulse shaping the GFDM subcarriers without offset QAM modulation or
orthogonal pulses is that the data symbols within a GFDM block interfere with each other. While the
FP7 ICT GA 318555
D3.3
zero-forcing receiver is an efficient tool to avoid this self-interference, it bears the drawback of noise
enhancement. Due to the discrete Gabor setting this noise enhancement heavily depends on the
system parameters: while the Balian-Low theorem prohibits efficient operation it can be shown that
for the discrete setting here in certain cases the Balian-Low theorem can be circumvented at least to
avoid sampling the DZT at zeros [MMF14]. Moreover, to avoid amplifying the noise in the reception
process, a Matched Filter (MF) receiver can be used to maximize the SNR for the individual
subcarriers, in combination with a subsequent Successive Interference Cancellation (SIC) stage to
remove the non-orthogonal parts in the signal, before detection. In GFDM, SIC is particularly
effective, because each subcarrier is well localized in frequency domain and only immediate
neighbors interfere with each other. Results have shown that MF-SIC can effectively remove self-
interference at the cost of a reasonable computational complexity [GAS13].
2.2.3 Multi-User Time-Reversal Space-Time Coding for GFDM
2.2.3.1 Time-Reversal Space-Time Coding

Time-Reversal Space-Time Coding (TR-STC) has been proposed by [Dha01] to allow the use of STC for
single carrier transmission over frequency-selective channels. The proposed approach operates on
two subsequent data blocks x i and x i 1 of length N which are separated by a CP. Their
corresponding discrete Fourier transforms are X ()
F x () , where F denotes the unitary Fourier
matrix. The transmit signal on both antennas for two subsequent time slots is given by
A n te n n a 1 A n te n n a 2
F
H H *
B lo c k i F X i
X i 1
B lo c k i 1
H H *
F X i 1
F X i
where i is an even number and () * is the conjugate complex operation. Note that the property
F H
X
*
i n
xi n

*
N

of the discrete Fourier transform reasons the name time-reversal space-time coding.
At the receiver, after removing the CP, the transmit signals appear circularly convolved with the CIR
h j , l , where h j , l contains the channel taps between the j th transmit and l th receive antenna, zero-
padded to the block length. Both received blocks are transformed to the frequency domain.
Accordingly, assuming the channel remains constant during the transmission of two subsequent
blocks, the received blocks in the frequency domain are given by
Y i ,l H 1,l X i H W 1,l
*
2 ,l
X i 1
Y i 1,l H 1,l X H W 2 ,l ,
*
i 1 2 ,l
X i
where H j ,l
d ia g ( H j ,i
) with H j ,i
F h j ,l . The received signals can be combined in the frequency
by
FP7 ICT GA 318555
D3.3
L

1
Xi H H 1,l Y i ,l H
* *
eq 2 ,l
Y i 1,l
l 1
L
(2.2.3.1.1)

1
X i 1 H H 1,l Y i 1,l H
* *
eq 2 ,l
Yi ,l
l 1
2 L
where H e q
*
H j ,l
H j ,l
and L is the number of receiving antennas. Finally, the estimates of
j 1 l 1
the transmitted blocks are acquired by inverse Fourier transform

H
x i F and x i 1 F X i 1 .
H
Xi
2.2.3.2 Multi-User STC-GFDMA
TR-STC can be directly applied to GFDM. Consider two data vectors d i that generate two
consecutive GFDM frames
xi A d i
The GFDM signals x i and x i 1 can be space-time encoded as described in Sec. 2.2.3.1 and (2.2.3.1.1)
can be used to recover the signals on the receiver side. Then, conventional GFDM ZF demodulation is
carried out by

d i B x i
In this section, we combine TR-STC GFDM with a FDMA technique to serve multiple users in one
GFDM system. The K subcarriers are equally divided between U users, i.e. each user allocates
K u K / U adjacent subcarriers. Clearly, bit loading and power allocation algorithms [PK11] can be
used to better distribute the resources between the users, however, this is out of scope of this
analysis. Fig. 2.2.4 depicts the block diagram of the proposed TR-STC-GFDMA system.
Figure 2.2.4: Simplified block diagram of the TR-STC-GFDMA system
(u )
Each user u generates a GFDM signal based on two successive data vectors d i , where elements
corresponding to non-allocated subcarriers are set to zero. The data is modulated by A for each
user and the blocks are space-time encoded as described in Section 2.2.3.1. In order to combat time
misalignment between users, in addition to the CP, a cyclic suffix (CS) is added to the blocks before
they are transmitted through independent frequency-selective fading channels. When using a non-
orthogonal transmit filter, adjacent subcarriers of two different users interfere with each other. Since
the channels for the users are independent, these boundary subcarriers cannot be equalized and
high ICI occurs. Therefore, one guard subcarrier is used between users to avoid mutual interference.
The guard subcarriers are unnecessary when an orthogonal pulse is used because, in this case, there
is no ICI.
FP7 ICT GA 318555
D3.3
When all users are synchronized with the BS clock rate and all signals arrive within the CP and CS
duration at the BS, space-time combining according to (2.2.3.1.1) can be carried out per user, where
only the users' allocated frequency samples are considered. However, CSI including misalignment
information needs to be available at the BS. Note that timing misalignment within the CP/CS length
only results in a phase rotation of the circulant channel and hence channel equalization can
compensate the misalignment. The CSI can be estimated at the BS by sending a separate pilot
sequence per user that is transmitted twice in the TR-STC-GFDMA codeword. The sequences are
modulated using an orthogonal pulse to avoid self-interference, with K u subcarriers per user and
M 1 subsymbols. The signals are time-reversal space-time encoded and CP and CS are appended
before transmission.
At the BS, the received signal in the frequency domain is given by

Pi , l P H 1 , l P H W i ,l
*
2 ,l
Pi 1 , l P H 1 , l P H W i 1,l
*
2 ,l
where P d ia g ( P ( u ) ) with P
(u )
containing the DFT of the pilot sequence of the u th user which
u 1
T T
[H
(1 ) (U ) T
is non-zero only at the allocated subcarriers and H j ,l j ,l
H j ,l
] denotes the frequency
response of the users' channels including time misalignment seen at the BS. Then, the CIR is
estimated by
1
H 1,l 1
2
P ( Pi , l Pi 1 , l ) ,
1
(2.2.3.2.1)
( Pi 1 , l Pi , l ) .
1 *
H 2 ,l 2
(P )
Note that the proposed scheme does not require all users to have two transmit antennas. Instead,
using only the first transmit antenna is equivalent to h 2 , l 0 . In this case, the transmitted signal can
still be recovered, however, according to (2.2.3.1.1) no transmit diversity gain is achieved. For
example, users with cheap devices or very good channel conditions wouldn't carry out the space-
time encoding but can still be correctly received by the BS.
2.3 UFMC
Universal filtered multicarrier (UFMC) can, roughly speaking, be interpreted as a generalization of

FBMC and filtered OFDM [SWC14]. While the former filters each subcarrier and the latter filters the
entire band, UFMC filters blocks of subcarriers, as depicted in Fig. 2.3.1. Subband-wise filtering is
motivated by the observation that time-frequency misalignments typically occur between entire
blocks of subcarriers (e.g. due to block-wise resource allocation of different uplink users).
Furthermore, as the filters are broader in frequency, they become shorter in time. This provides a
good support for communication in short bursts. A summary on UFMC-based 5G air interface design
advantages and available results is provided by [WSC14]. Note that UFMC is also called UF-OFDM, as
with filter length L=1, conventional OFDM is a subset of UF-OFDM. Matlab code of the basic UFMC
transceiver can be downloaded from [SW].
The time-domain transmit vector x k for a particular multi-carrier symbol of user k is the
superposition of the sub-band-wise filtered components, with filter length L and FFT length N, and
can be represented by
FP7 ICT GA 318555
D3.3
B
xk F ik V ik s ik (2.3.1)
( N L 1) 1 i 1 ( N L 1 ) N N n n
i i
1
For the i-th subband (1 i B), the ni complex QAM symbols are transformed to time domain by the
tall IDFT-matrix V ik . Its m-th row, for subcarrier index n consists of the elements Vik(m, n) = exp
(j2(m 1)n/N). F ik is a Toeplitz matrix, composed of the FIR filter impulse response f ik , performing
the linear convolution. f ik can be designed according to propagation conditions and time-frequency
offset requirements (see section 4.1.2 and [WWS+14]). For L = 1, UF-OFDM converges to (non-CP-)
OFDM.
There are several receiver implementations possible for UFMC. A very efficient approach [CWS14] is
to zero-pad the N + L 1 receive samples of a multi-carrier symbol to the next larger power of two
(or efficient FFT implementation length). Each second value in frequency represents a subcarrier
output, which can be treated by a 1-tap scalar equalizer in order to account for the transmit filter
frequency response. This makes reception of UFMC signals almost as low complex as CP-OFDM.
Figure 2.3.1 Basic UFMC transceiver chain
FP7 ICT GA 318555
D3.3
2.3.1 UFMC Frequency- and Time Domain Properties

The spectrum of one subband is depicted in figure 2.3.2. The OFDM drawbacks of high side-lobe
levels can be removed with UFMC. This makes UFMC suitable for fragmented spectrum and increases
robustness against all sources of inter-carrier interference (ICI). This ICI robustness was shown for
uplink coordinated multi-point (CoMP) joint reception in the presence of carrier frequency offsets of
multiple users in [VWS+14]. UFMC thus becomes an enabling waveform for the Unified Frame
Structure (see section 3.1), supporting various traffic types and devices, including time-frequency
misalignments [SW14].
-10
Rel. power [dB]
-20
-30
-40
-50
-60
0 20 40 60 80 100 120 140
Frequency spacing in subcarrier steps
Figure 2.3.2 UFMC spectrum (green) compared to OFDM spectrum (black) for one subband of 12 subcarriers
with QPSK modulation
As long as the same length subband-filter is used, the subcarriers in UFMC are fully orthogonal in the
complex domain. This means, in contrast to FBMC-SMT (described in the subsequent section), which
uses offset-QAM, UFMC can use QAM modulation symbols. This complex orthogonality eases the
usage of MIMO and CoMP, supports complex-valued pilot sequences and allows re-using all the
knowledge gained from CP-OFDM.
OFDM, LCP = 79
0.06
0.04
0.02
Re(x k)
-0.02
-0.04
0 500 1000 1500 2000
time index m
UFMC, L = 80, SLA = 60

0.06
0.04
0.02
Re(x k)
-0.02
-0.04
0 500 1000 1500 2000
time index m
Figure 2.3.3 Time domain characteristic of UFMC (bottom) compared to CP-OFDM (top), depicting amplitude
real part of a single subcarrier. The soft symbol transition in UFMC (with light blue background) provides a soft
protection against inter-symbol interference
FP7 ICT GA 318555
D3.3
In time domain, UFMC in its default form is characterized by soft transitions between different multi-
carrier symbols. Subsequent multi-carrier symbols, characterized by (2.3.1), are non-overlapping.
This means the enveloping amplitude of the subcarriers, ramps up, stays constant for the symbol
body and ramps down again, as shown in figure 2.3.3. This provides a soft inter-symbol-interference
protection (ISI). E.g. the MSE caused by the delay spread of a Vehicular A channel, when UFMC
operates with LTE-like parameters is far below -40dB (when the filter length L = LCP + 1 accounts for
the CP length LCP in LTE). This allows for the scalar symbol-wise equalization and omitting the cyclic
prefix. See section 4.1.2 for results on time-frequency efficiency in short burst settings, comparing
CP-OFDM, FBMC and UFMC.
2.3.2 Channel estimation and equalization

As already discussed in [D3.2], channel estimation and equalization are extremely simple for UFMC.
All the know-how generated for OFDM can be directly reused.
UFMC channel estimation can be executed in the frequency domain. After the FFT, scalar per-
subcarrier processing in the frequency domain can be done, which is low complex and can build upon
OFDM knowledge. First a raw channel estimate for a single resource element can be computed based
on the known pilot symbol SPilot(n), known filter frequency response FFD(n) and observed subcarrier
received value Ysingle(n) as :
H CTF
( n ) Y single (n) F FD ( n ) S Pilot ( n ) (2.3.2)
Compared to OFDM, the impact of the filtering is additionally taken into account.
After having achieved the raw channel estimate (2.3.2), every subsequent processing, as known from
OFDM can be applied. E.g. two-dimensional Wiener filtering in time- and frequency dimension will be
identical to OFDM.
Using the estimated channel, a per-subcarrier scalar equalizer can be applied, performing an
element-wise multiplication of equalizing vector q and frequency response vector Ysingle
s q Y single (2.3.3)
where the circular symbol represents the Hadamard-product, carrying out the element-wise
multiplication.
The equalizing vector takes care for all phase rotations caused by the frequency responses of the
respective per-subband filters, including the filter delay phase shifts, as well as the frequency
response of the channel HCTF, the so-called channel transfer function. Note that this equalization is
similar to OFDM (with additional compensation of the filter).
2.3.3 Native MIMO/CoMP support

The extension to MIMO for UFMC is straightforward. The UFMC waveform decomposes the
frequency-domain channel into orthogonal narrowband sub-channels, as in OFDM. In order to
achieve the MIMO capacity, known approaches can be applied: A singular value decomposition of
the MIMO channel matrix, followed by waterfilling. Practical usage of MIMO is simple: Complex
precoding and QAM modulation are supported, as in OFDM. Results for multi-user uplink distributed
MIMO, thus coordinated multi-point (CoMP) joint reception are provided in section 4.1.2.
FP7 ICT GA 318555
D3.3
2.3.4 Support for multi-user sounding and multiplex of small control resource element groups
In order to efficiently support control signaling, small groups of resource elements have to be used
per user to carry information like channel state information and ACK/NACK. In LTE e.g. the PUCCH is
responsible for such tasks.
In uplink sounding, the base station gathers channel state information over a certain frequency range
for multiple users. This requires either a comb-like structure for different users or a superposition of
different sequences. LTE(-A) based on CP-OFDM supports both, using complex-valued Zadoff-Chu
sequences.
For non-orthogonal waveforms the multiplexing of multiple users for sounding and control may
impose difficulties. Different uplink users are received via different propagation channels. In order to
avoid inter-user interference either overhead by guard subcarriers/time will be introduced or
receiver complexity will strongly increase.
For UFMC this is not the case: As the filtered UFMC multi-carrier symbols are non-overlapping in time
and well localized in frequency, this waveform very efficiently supports small groups of resource
elements carrying control information. While offset-QAM based waveforms only use real-valued
sequences for sounding etc., complex-valued sequences are fully supported in UFMC, so the
sequence-space dimensionality for multi-cell and multi-user support is much larger compared to
FBMC.
2.3.5 Adaptivity Potential

The basic signal generation for UFMC is given by (2.3.1). This signal generation can be expanded to
cover a single-carrier DFT-precoded variant of UFMC and/or to precompensate for the pass-band
filter response, denoted by
B
xk F ik V ik P ik D ik s ik .
( N L 1) 1 i 1 ( N L 1 ) N N n n
i i
ni n i
ni n i
1
(2.3.4)
Here, D ik is a diagonal matrix which e.g. compensates the pass-band frequency domain complex
filter response to generate a frequency flat subband with no phase rotations. The advantage is that
the receiver does not need to know the actual used UFMC filter coefficients which allows for further
adaptation potential without signalling overhead. Additionally, it guarantees equal transmitted
power per subcarrier.
P ik is a precoding matrix. When this precoding is the DFT matrix, we generate a single carrier FDMA
signal (similar to LTE uplink), based on the UFMC waveform instead the CP-OFDM waveform. This
allows for a peak to average power ration (PAPR) reduction with similar impact as in the LTE uplink
with DFT-precoded OFDM, but preserving the better spectral properties of UFMC. (In this notation
P ik covers a single subband, but this can be easily changed to e.g. the entire contiguous allocation
range of a user for further PAPR reduction.)
When the UFMC subband filter f ik contains zeros at the beginning or end, we have generated a zero
prefix or postfix signal, providing additional protection against delay spreads and/or timing offsets.
FP7 ICT GA 318555
D3.3
Note that all this adaptations in conjunction with FDMA allow for user-individual adaptation of UFMC
parameters. The conditions of each user can be individually taken into account: Delay spread, timing
and frequency offsets, and PAPR requirements. For those conditions, the parameter set for subband
i, assigned to a particular user, can be designed appropriately.
The SC-FDMA incarnation of UFMC, generated by the DFT-precoding matrix P ik , can additionally be
adjusted to delay spreads in a user-specific way with near zero tails. In this case we apply a technique
known in the OFDM community as zero tail DFT-spread OFDM [BTS+13] to UFMC: Users with high
delay spreads are configured to set a few DFT-precoder inputs (first and last elements of s ik ) to zero
at the allocation edges, which leads to strong signal level reduction at beginning an end of the
symbol and thus increases protection against delay spreads.
2.4 FBMC
In this section, bold letters denote vectors and matrices. Upper-case and lower-case letters denote
frequency domain and time domain variables respectively
F stands for the N N -DFT (Discrete Fourier transform) matrix defined as:
1 1 1 1
2 N 1
1 wN wN wN
1
F =
N N 1 2( N 1) ( N 1)( N 1)
1 wN wN wN

2
j
where w N = e N
. Matlab notation was used to index the matrix. Therefore A = B (:,1 : U ) means
that A is built with the first U columns and all the rows of B .
2.4.1 FBMC principles

A multicarrier system can be described by a synthesis-analysis filter bank, i.e. a transmultiplexer
structure. The synthesis filter bank is composed of all the parallel transmit filters and the analysis
filter bank consists in all the matched receive filters, as shown in Figure 2.4.1 where () and
() are respectively the transmit and receive prototype filters. For subcarrier , the filter is the
prototype filter phase shifted by 2 . This phase shift in the time domain implies a frequency
shift of in the frequency domain. In this figure, the data signal is defined by Eq. (2.4.1):

sk t sk n t nT (2.4.1)
n
with [] the data symbols for subcarrier , the symbol period, the symbol number and Nc the
number of subchannels.
The most widely used multicarrier technique is CP-OFDM, based on the use of inverse and forward
DFT for the analysis and the synthesis filter banks. The prototype filter is a rectangular window
whose size is equal to the Fourier Transform. At the receiver, perfect signal recovery is possible
under ideal channel conditions thanks to the orthogonality of the subchannel filters. Nevertheless
FP7 ICT GA 318555
D3.3
under real multipath channels a data rate loss is induced by the mandatory use of a Cyclic Prefix (CP),
longer than the impulse response of the channel. With FBMC, the CP can be removed and subcarriers
can be better localized, thanks to more advanced prototype filter design.
s0 t pTx t e j 2 f0t pRx t e j 2 f0t s0 n
s1 t pTx t e j 2 f1t pRx t e j 2 f1t s1 n

+ Channel
sNc 1 t pTx t e j 2 f Nc1t pRx t e j 2 f Nc1t sNc 1 n
transmitter receiver
Figure 2.4.1: Block diagram of a multicarrier transceiver
The FBMC prototype filter can be designed in many ways, trying to satisfy different constraints. In
general, it is chosen to be:
- complex modulated for good spectral efficiency
- uniform to equally divide the available channel bandwidth
- with finite Impulse Response for ease of design and implementation
- orthogonal, to have a single prototype filter
- with Nearly Perfect Reconstruction (NPR) : certain amount of filter bank distortions can be
tolerated as long as they are negligible compared to those caused by the transmission
channel
In this document the prototype filter is designed using the frequency sampling technique. This
technique provides the advantage of reducing the number of filter coefficients. In other words, the
prototype filter coefficients should be given using a closed-form representation that includes only a
few adjustable design parameters. The coefficients of the prototype filter for an overlapping factor K
equal to 4 are [Bel10]:
P0 :3 1, 0 .9 7 1 9 5 9 8 3 ,1 2 , 1 P1
2
(2.4.2)

The KNc-1 length time response of this filter is computed thanks to:
K 1
2 k
p m P0 2 1 m 1 , m 0 : K N c 2
k
Pk c o s (2.4.3)
k 1 KNc
The stopband attenuation exceeds 60 dB for the frequency range above 10 subcarrier spacings
(Figure 2.4.2).
Figure 2.4.2: FBMC - filters for subcarriers 0 (blue) and 1 (red), = , =
FP7 ICT GA 318555
D3.3
As shown in Figure 2.4.2, adjacent carriers significantly overlap. In order to keep adjacent carriers
and symbols orthogonal, real and pure imaginary values alternate on carriers and on symbols at the
transmitter side. This so-called OQAM (Offset QAM) modulation implies a rate loss of a factor of 2.
This efficiency loss of OQAM modulation is compensated by doubling the symbol period . is
frequently called the overlapping factor: indeed the symbol period is /2 and the symbol length is
1 samples; each FBMC symbol at the channel input is then overlapped with 2(2 1) other
FBMC symbols.
The transmitter of FBMC can be represented by Figure 2.4.3, with the filtering operation (block
frequency spreading) done in the frequency domain. In this figure:
1 is the vector containing the data to transmit for the th FBMC symbol.
= , 1 is the vector of data for the th FBMC symbol filtered in the
frequency domain.
is the matrix of filtering vectors given by Eq. (2.4.4), with 121 , the
filtering vector, i.e. the frequency response of the filter given by Eq. (2.4.2):
(2.4.4)
1 is the vector of data for the th FBMC in the time domain.

The IFFT has a size of samples. The transmit signal is composed of the overlapping of symbols
with a factor of /2.
Figure 2.4.3: FBMC transmitter with filtering in the frequency domain
The frequency spreading operation is further described by Figure 2.4.4 where each carrier of is
spread on 2 1 carriers on . Here = 4. As can be seen from Figure 2.4.3, FBMC symbols
overlap in the time domain and as Figure 2.4.4 shows, adjacent carriers in the vector significantly
overlap in the vector .
Figure 2.4.4: Frequency spreading
FP7 ICT GA 318555
D3.3
The FBMC waveform with its spectrally well shaped prototype filters and overlapped time symbols
has some inherent features which makes it a natural choice for some of the anticipated 5G
application scenarios. First of all it does not require a cyclic prefix; intrinsic Inter Symbol Interference
(ISI) and ISI generated by multipath channel are handled by O-QAM modulation that ensures
orthogonality between adjacent symbols. The FBMC waveform furthermore has an almost perfect
separation of frequency subbands without the need for strict synchronization. Consequently its
properties make it especially suited for fragmented spectrum and Coordinated Multi Point (CoMP)
Transmission/Reception.
2.4.2 FBMC receiver

The dual operation of the overlap-and-sum operation of the transmitter is a sliding window in the
time domain at the receiver that selects KN-points every N/2 samples. A FFT is then applied every
block of KN selected points. Figure 2.4.5 illustrates the principle of the Frequency-Spreading FBMC
(FS-FBMC) receiver. When a FS-FBMC receiver architecture is considered, frequency-domain
equalization is performed after the FFT but before applying the matched filter. As a consequence,
channel estimation should also be realized before filtering. One advantage of this architecture is that
frequency domain time synchronization may be performed independently of the position of the FFT
[Bel12]. This is realized by combining timing synchronization with channel equalization. Another main
benefit of FS-FBMC is that channel equalization may be limited to a one-tap complex-multiply
operation while still sustaining significant tolerance to channel impulse response delay spread, even
compared to conventional PPN-FBMC architecture as shown in [BDN14]. Because the size of the FFT
is K-times larger than the multicarrier symbol time period, the signal at the output of the FFT is
oversampled by a factor of K with regards to the carrier spacing. This property gives a significant
advantage to FS-FBMC when the channel is exhibiting large delay spread.
The reader may refer to [D3.2] for further details on time synchronization, carrier frequency offset
compensation, channel estimation and equalizer algorithm.
Figure 2.4.5: FS-FBMC receiver principles

FP7 ICT GA 318555
D3.3
2.5 BFDM
A major new approach in 5GNOW for efficiently supporting sporadic traffic in 5G is to use an
extended physical layer random access channel (PRACH) which achieves device acquisition and
(typically small) payload transmission "in one shot" [WJK+14, KWJ+14, WKJ15, WGA15, WKJ+15], for
the details please see section 3.4.1. For this concept we adopt a waveform design approach based on
bi-orthogonal frequency division multiplexing (BFDM) for the PRACH signal (pulse shaped PRACH).
Similar to orthogonal OFDM, the underlying principle is to transmit the symbols according to a set
of shifted pulses on time-frequency lattice points (, ), where is the time shift period and is
the frequency shift period and , . However, as stated in [KM98], the only requirement of
perfect symbol reconstruction is that the set of transmit pulses , and the set of receive pulses ,
form bi-orthogonal (Riesz-) bases which is a weaker form of orthogonality and which is possible only
if the time-frequency product is greater than one. Hence, in the BFDM approach, we replace the
orthogonality of the set of transmit and receive pulses with bi-orthogonality. In particular, time--
frequency representations of transmit and receive pulses are pairwise (not individually) orthogonal.
Thus, there is more flexibility in designing a transmit prototype, e.g., in terms of side-lobe
suppression.
2.5.1 Transmitter and receiver Structure

In the following, let denote the sampling period, which is equal to = 1/ , with being the
sampling frequency. In the following discrete model, we let all time indices be multiples of and
frequency indices be multiples of . Furthermore, we use $N$ to denote the discrete counterpart of
the symbol duration and submit symbols. Let be the FFT length. Note, for some numerical
reasons must divide . We choose TF=1.25 to be compliant with LTE.
For the pulse shaped PRACH, additional processing is needed, compared to standard OFDM. In
contrast to standard processing, we process more than one symbol interval, even if we use only one
symbol to carry the preamble. We refer to [SMH02] for implementation details. A pulse is used to
shape the spectrum of the preamble signal, e.g., to allow the use of PRACH guard bands with
acceptable interference. Let be the length of the pulse . We extend the output signal [] after
the IFFT stage by repeating it and taking modulo to get the same length as the pulse . Given
symbols, we stack each symbol [] as rows in a matrix
0 []
= 1 []

( 1 [])

Each of these vectors is point-wise multiplied by the shifted pulse and superimposed by overlap
and add, such that we get the base band pulse shaped PRACH transmit signal
1

[] = [][ ].
=0
In greater detail, this can be also written as
2

[] = , [ ] ,
where , is the Fourier transformed ZC-sequence of length at the th symbol and th
subcarrier, is the guard band subcarriers occupied by messages, and is an amplitude scaling
factor for customizing the transmit power.
FP7 ICT GA 318555
D3.3
The BFDM-based transmitter structure is illustrated schematically in Figure 2.5.1.
Figure 2.5.1 BFDM based pulse shaped PRACH transmitter
The only difference to the standard PRACH receiver is the processing before the FFT. In standard
PRACH processing, the cyclic prefix is first removed from the received signal [] and then the FFT
is performed. In the pulse shaped PRACH, an operation to invert the (transmitter side) pulse shaping
has to be carried out first. To be more precise, first the symbols of the received signal [] are
arranged as row vectors in matrix
0 []
= 1 []

(1 [])
Second, each row is point-wise multiplied by the shifted bi-orthogonal pulse , such that we have

[] = [][ ].

Subsequently, we perform a kind of pre-aliasing operation to each windowed []
/ 1

[] = [ ],
=0
such that we obtain the Fourier transformed preamble sequence at the th symbol and th
subcarrier after the FFT operation
1 2

, = [] .
=0
Although we do not employ a cyclic prefix as in standard PRACH, the time--frequency product of
= 1.25 allows the signal to have temporal and frequency guard regions as well. This time-
frequency guard regions and the overlapping of the pulses evoke the received signal to be cyclo-
stationary [Bol01], which gives the same benefit as the cyclo-stationarity made by cyclic prefix.
Furthermore it is also shown in [Bol01] that the bi-orthogonality condition of the pulses is sufficient
for the cyclo-stationarity and makes it possible to estimate the symbol timing offset from its
correlation function.
The BFDM-based PRACH receiver is illustrated schematically in Figure 2.5.2.
FP7 ICT GA 318555
D3.3
Figure 2.5.2 BFDM based pulse shaped PRACH receiver
2.5.2 Pulse design
2.5.2.1 Spline waveforms

As mentioned before, the used pulses and play a key role and must therefore be carefully
designed. Since we consider here the BFDM approach, we setup the transmit pulse according to
system requirements and compute from the receive pulse as the canonical dual (bi-orthogonal)
pulse. For this computation we follow here the method already used, for example, in [JW07] (see
also the further references cited therein). Briefly explained, bi-orthogonality in a stable sense means
that should generate a Gabor Riesz basis and generates the corresponding dual Riesz basis. From
the Ron--Shen duality principle [RS97] follows that has the desired property if it generates on the
so called adjoint time--frequency lattice a Gabor (Weyl--Heisenberg) frame which is dual to the frame
generated by . However, this can be achieved with the 1--trick explained in [Dau92].
As a rough and well--known guideline for well-conditioning of this procedure, the ratio of the time
and frequency pulse widths (variances) and should be approximately matched to the time
frequency grid ratio

, (2.4.2)

which should also be in the order of the channel's dispersion ratio [JW07]. However, here we
consider only the first part (2.4.2) of this rule since we focus on a design being close to the
conventional LTE PUSCH and PRACH. We propose to construct the pulse based on the -splines in
the frequency domain. -splines have been investigated in the Gabor (Weyl--Heisenberg) setting for
example in [Pre99]. The main reason for using the -spline pulses is that convolution of such pulses
have excellent tail properties with respect to the 1 -norm, which is beneficial with respect to the
overlap of PRACH to the PUSCH symbols. We also believe that they trade off well the time offset for
the frequency offset performance degradation but this is part of further on-going investigations and
beyond the conceptional approach here.
Because of its fast decay in time, we choose in [KWJ+14] a second order -spline (the ''tent''--
function) in frequency domain given by
2 () = 1 () 1 (), where
1 () 1 1 ()
[ , ]
22
FP7 ICT GA 318555
D3.3
It has been shown in [Pre99] that 2 () generates a Gabor frame for the (, )-grid (translating 2
1
on and its Fourier transform on ) if (due to its compact support) < 2 and 2 and fails to be
frame in the region:
{ 2, > 0} { > 0,1 < }.
Recall, that by Ron--Shen duality [RS97] it follows that the same pulse prototype 2 () generates a
1 1
Riesz basis on the adjoint ( , ) -grid. In our setting we will effectively translate the frequency
1 3 1 1 5
domain pulse 2 () by half of its support which corresponds to
= 2 and we will use
=4=
6
1.25 (see here also Table 2.4.1) such that = 5. It follows therefore that our operation point
6 2
(, ) = ( , )is not in any of two explicit (, )-regions given above. But for 1.1 1.95 a
5 3
further estimate has been computed explicitly for 2 () [Table 2.3 on p.560, Pre99] ensuring the
1 1
Gabor frame property up to . Finally, we like to mention that for the dual prototypes
2
can be expressed again as finite linear combinations of -splines, i.e. explicit formulas exists in
[Lau09].
However, in practice has to be of finite duration, i.e. the transmit pulse in time domain will be
smoothly truncated
sin() 2
() = ( ) (), (2.4.3)
where is chosen equal to . Theoretically, a (smooth) truncation in (2.4.3) would imply again a
limitation on the maximal frequency spacing [CKK12]. Although the finite setting is used in our
application, the frame condition (and therefore the Riesz-basis condition) is a desired feature since it
will asymptotically ensure the stability of the computation of the dual pulse and its smoothness
properties. To observe the pulse's properties regarding time-frequency distortions we depict in
Figure 2.4.3 the discrete cross-ambiguity function between pulse and which is given as:
(, ) = [] [ ] 2 .

Figure 2.5.3 The cross-ambiguity function (, )
FP7 ICT GA 318555
D3.3
It can be observed, that the value at the neighbour symbol is already far below 103. Obviously, the
bi-orthogonality condition states (, ) = ,0 ,0 and ensures perfect symbol recovery in the
absence of channel and noise. However, the sensibility with respect time-frequency distortions is
related to the slope shape of around the grid points. Depending on the loading strategies for
these grid points it is possible to obtain numerically performance estimates using, for example, the
integration methods presented in [JW07].
2.5.2.2 Trapezoidal pulses

This pulse is obviously positive and real for all times. Moreover the performance with respect to
offsets in time and delay curves is superior to OFDM pulses. However, the pulse lacks some
additional parameters for the adaption to different settings. For this we define the trapezoidal family
given by
(+1) (1)
2 sin( ) sin( )
() = 2
2
2
,
(1)
of which the Fourier transform is depicted in Fig. 2.5.5. This family is controlled by the parameter L
(and bandwidth B) which contains the special cases: L = 0 (triangle) L = 1 (sinc) L > 1 (trapezoidal).
Our specific choice is guided by the following intuition: From the results in [KWJ+14] we conclude
that the triangle pulse belongs to the non-orthogonal pulse family with excellent localization
properties. This is obvious in the frequency domain; but to see this in the time domain we consider
the 1 -norm. Surprisingly, for the triangle kernel, this norm is actually independent and unity for all B
since:

1 = 2 ( ) = 1.
2 2 2
Here, we introduced the sinc pulse sinc(At) := sin(At)/At, A > 0. It is easy to prove that no pulse with
f(0) = 1 can fall below this value. Hence, we can argue that for any pulse 1 > 1 (after proper
normalization) and the 1 -norm measures the (inverse) distance to the sinc pulse for which clearly
1 = holds. We also argue that our goal is to shift the triangle closer to transmit
orthogonality (i.e. sinc) but at the same time not lose the favorable properties of the ambiguity
function of the triangle pulse. This is achieved by the trapezoidal family which has close to optimal
1 -norm behavior: in [WG15] we show that indeed
1 2 (),
where > 2 () > 1 and for which some exemplary values are listed in Table .
Table 2.5.1
Exemplary values of (): The values quickly converge to 1
(triangle pulse)
2 ()
1
1.5 1.68
2 2
2.5 1.25
. .
. .
. .
1
FP7 ICT GA 318555
D3.3
Fig 2.5.4: Frequency domain response of trapezoidal filter

Let us re-write the trapezoidal pulse in the following more convenient form by incorporating
parameters as
+
, () = sinc ( ) sinc ( ),
2 2
whose frequency domain shape is sketched in Fig. 2.5.4. In order to be compliant to 4G, we chose
= 1.25 and we set = in the pulse definitions. We consider two types of trapezoidal pulses.
Type I refers to transmit pulses with constant bandwidth 2, i.e. = 1, and the inner flat part over
the frequency interval [, ]. Type II pulses have an inner flat part over the fixed interval

[ , ], i.e. = 0.5, and variable bandwidth [, ]. The time and frequency domain shape of
2 2
these pulses for both the transmit and dual receive pulses are depicted in Fig. 2.5.5 for the
parameters of interest in the simulations. In particular, the bold curve refers to a type II pulse with
= 0.55 which demonstrates the best performance among the considered pulses.
Figure 2.5.4. Pulse shapes considered in the simulations: Time domain for transmit (a) and receive (c) pulse
shapes as well as frequency domain for transmit (b) and receive (d) pulse shapes. Type I with = , = .
(black) and Type II with = . , = . (green) and 0.75 (red). Triange pulse (blue)
FP7 ICT GA 318555
D3.3
2.5.3 Synchronization and Equalization

Due to the Random Access Scenario that we consider, exact synchronization is not needed. In fact, a
major design target is to enable asynchronous access and data transmission. Nevertheless, we have
at least a course synchronization due to the reference signals of the base station in downlink. For the
discussion of channel estimation and equalization, we first summarize the user detection procedure
in PRACH.
User detection in PRACH is based on suitably chosen preamble sequences, which are generated as
follows. The preamble is constructed from a Zadoff-Chu (ZC) sequence [Chu72], which is defined as
( + 1)
[] = exp { } , 0 1,

where is the root index and is the length of the sequence. We consider here the case of
contention-based RACH, where every user wanting to send a preamble chooses a signature randomly
from the set of available signatures = {1, ,64 }, with being a given number of reserved
signatures for contention free RACH. Every element of is assigned to index (, ), such that the
preamble for each user is obtained by cyclic shifting the th Zadoff-Chu sequence according to
, [] = [( + ) ],

where = 1, , is the cyclic shift index, is the cyclic shift size and is the preamble

length which is fixed for all users. Since there can only exist preambles that can be generated

from the root , the assignment from to (, ) depends on and the size of set .
The signatures of different users can be detected as follows. Given the received signal (2.4.1), the
PRACH receiver observes the fraction that lies in the PRACH region to obtain the preamble. The
receiver stores all available Zadoff--Chu roots as a reference. These root sequences are transformed
to frequency domain and each of them is multiplied with the received preamble. The result is
[] = [] [],
where [] is the received preamble and [] is the th ZC sequence in frequency domain
respectively. Using the convolution property of the Fourier transform it is easy to show that [] is
equal to the inverse Fourier transform of any cross correlation function [] at lag . Because the
preamble is constructed by cyclic shifting the Zadoff--Chu sequence, ideally we can detect the
signature by observing a peak from the power delay profile, given by
2
| []|2 = | 1
=0 [ + ] []| . (2.4.4)

Let = be the maximum number of preambles which can be generated from one root. Then

64
the number of roots that we require to generate 64 preambles will be = . The
signature and the delay of user , denoted by and , respectively, are obtained by the following
operation

= + , 0 ,

= ( ) ,

where is the location of the largest peak in (2.4.4).
The question remains how to obtain some estimation for the channel. Assume the received preamble
signal can be written as
= .

Thereby, is a diagonal matrix constructed from the coefficients of the Fourier transformed
preamble and = ( , ). The matrix is a FFT-matrix, the set {1, , } contains
FP7 ICT GA 318555
D3.3
the indices of the first columns, and = {1, , } contains the indices of the central rows
of . Furthermore, is the length of the subframe without CP and guard interval, and we assume a
maximum length of the channel .
For simplicity, we consider simple least-squares channel estimation, i.e., we have to solve the
estimation (normal) equation = . To handle cases where is ill-conditioned, we use
Tikhonov regularization. This popular method replaces the general problem of minx 2 by
min 2 + 2 , with the regularization matrix . In particular, in place of the pseudo-
inverse, we use
= ( + )1 ,
where is a multiple of the identity matrix. The idea is, that the estimated channel is also valid for
subcarriers that are adjacent to the region for which we actually estimate the channel. Numerical
experiments indicate that the estimator is an unbiased estimator (with MSE smaller then 104) for
up to 200 subcarriers outside the region .
FP7 ICT GA 318555
D3.3
3 5GNOW Frame Structure and Multiple Access
3.1 Unified Frame Structure
In order to suit the future needs of very heterogeneous service and device classes, a 5G approach
must be able to efficiently support different traffic types, which all have to be part of future wireless
cellular systems. Our vision of a unified frame structure concept, depicted by Fig. 3.1.1, aims to
handle the large set of requirements in a single 5G system.
Layer
Type III and Type IV
Time
Type II
Type I
Frequency
Figure 3.1.1 5GNOW Unified frame structure concept
Here, different colors represent time-frequency resource elements of different traffic types and
different classes of synchronicity. A filtered multicarrier approach will enable the mix of synchronous
and asynchronous traffic. Radio resource control can adjust the assigned bandwidth semi-statically
based on the respective service loads. The third dimension in this Unified Frame Structure is the
usage of multiple superimposed signal layers. The principle of interleave-division multiple access
(IDMA) [PLK06] is a very appealing approach to generating these signal layers, and an elegant
receiver and coding concept for it. High-volume broadband traffic (Type I), typically human-initiated,
will operate as in LTE-A with synchronicity, whenever possible, using scheduled access. At cell edges,
with coordinated multi-point (CoMP) transmission and reception, it is not always possible to
establish synchronicity to all cells, so the system has to operate with relaxed synchronicity
requirements (Type II traffic). For high volume data applications in those cell areas (type II), a multi-
cell multi-user transceiver concept is required.
For sporadic small packet services, as occurring in e.g., MTC, the general relaxation of time-frequency
alignment reduces signaling overhead and battery consumption (Type III traffic). Multiple signal
layers may superimpose, which is handled by advanced multi-user / multi-cell receivers. A
contention-based access technique is attractive, saving overhead by dropping the strict synchronism
FP7 ICT GA 318555
D3.3
requirement. For low-end sensor devices, it may be helpful to spread the transmission over a longer
time and to allow completely asynchronous transmission (Type IV) traffic. For such sensor-type traffic
it can be shown that, from an energy-efficiency perspective, it is beneficial to stretch the
transmissions in time by spreading. This additional signal layer, again, can be handled by an IDMA-
like approach. The unified frame concept shall be main part of the standardization processes to be
initiated in the future.
In that context, exemplary multiple access schemes are defined as follows:

Dynamic, channel adaptive resource scheduling for traffic Type I using standard resource
scheduling mechanisms.
Semi-static/persistent scheduling for traffic Type II. From MAC point of view it is necessary to
decide on the amount of resources allocated for this type of traffic, since schedulers will not
adapt to specific parts of the frequency (may also be used for high speed terminals).
One shot transmission (low amount of data and pilots with contention-like based
approaches for random access (Type III and IV) enabling payload transmission in physical
layer random access channel (PRACH).
Notably, traffic types II and III rely on open-loop synchronization. The device listens to the downlink
and synchronizes itself coarsely, based on synchronization channel and/or reference symbols, similar
to 4G systems. Furthermore, the devices may apply some autonomously derived timing advance
which we call autonomous timing advance (see section 3.3), relevant particularly for MTC.
3.2 Layering
In the pioneer work [PLK06], the Interleaved-Division Multiple Access (IDMA) was originally proposed
by Li Ping et al., which was regarded as a novel multiple access approach. The authors claim that
IDMA can simultaneously accommodate, e.g. 40 ~ 100 users, by exploiting user-specific interleaver as
user signature and simply concatenating a repetition code with the Forward Error Correction (FEC)
code, to yield multiple low-rate data flows.
The IDMA receiver, denoted as Elementary Signal Estimator (ESE), turns out to be simple and
effective, as well, by collecting the statistics of received data streams, and involving Parallel
Interference Cancellation (PIC) to achieve the convergence with only a couple of iterations.
3.2.1 A Brief Review of IDMA

In this section, let us firstly have a brief review of IDMA concept. In Fig. 3.2.1, a block diagram of the
IDMA transmitter and receiver is illustrated. For a particular user n, the data bits dn are encoded,
denoted as coded bits cn. Each user is allocated a specific interleaver n . Hence, the conventional
IDMA system equation in a scalar form can be presented as
FP7 ICT GA 318555
D3.3
Figure 3.2.1: IDMA transmitter and receiver structure
y h n rn h l rl z (3.2.1)
ln
where rn denotes the QAM symbol of user n after user-specific interleaving, hn denotes the channel
transfer function, and z models Additive White Gaussian Noise (AWGN) with power spectral density
N0. At the receiver side, the ESE generates the Log-Likelihood-Ratio (LLR) for the m-th bit of the
transmit symbol rn with

2
y hn a
a A
0
exp( )
N N0
( rn ) ln
m
( out ) (m )
L ESE I
(3.2.2)
,n
y hn a
2

1
exp( )
a A m
N I
N0
0 1
where A m and A m denote the subsets of the QAM constellation candidates, whose m-th bit is 0 and
1, respectively. Notation N I
denotes the variance of the interference from other IDMA users.
Further, it holds,
y y h l E rl y h l Q ( L DEC,
( out ) ( out )
l ,1
, , L DEC, l ,M
) (3.2.3)
ln ln
where operator Q ( ) represents a soft QAM-mapper for symbol rn based on the LLRs from the
1
decoder. For the output of ESE, the user-specific de-interleaver n and FEC decoder basically
collect the diversity and coding gain, which can significantly improve the reliability of cancellation in
FP7 ICT GA 318555
D3.3
equation (3.2.3). Furthermore, the ESE performance (3.2.2) can be as well improved, by introducing
several IDMA iterations. The user signal layers 1 to n can be successfully separated even with, e.g. 5
IDMA iterations.
Figure 3.2.2: UFMC-IDMA transmitter and receiver structure
3.2.2 Involving IDMA in UFMC

In previous deliverables [D3.1] [D3.2], Universal Filtered Multi-Carrier (UFMC) [VWS13] technique is
proposed as an alternative solution for conventional CP-OFDM. Due to its close relatedness to CP-
OFDM, UFMC is also referred to as UF-OFDM. Comparing to CP-OFDM, UFMC technique is able to
FP7 ICT GA 318555
D3.3
provide better waveform spectral efficiency, lower spectral side lobe level, higher robustness for
tolerating time and frequency offsets and Inter-Carrier Interference (ICI), and reduced signaling
overhead [WSC14].
In [D3.2] and [CWS14], we introduce the IDMA concept to an UFMC system, as illustrated in Fig.
3.2.2. Further, a very simplified but fair dual-user scenario is established, in order to compare the
UFMC-IDMA scheme to the OFDM-IDMA scheme, in presence of relative delays. The numerical
results illustrate that UFMC-IDMA turns out to be the best solution at low rate, and can effectively
combat against misalignments, thus better to support the Relaxed Synchronicity (RS).
3.3 Relaxed Synchronization Approaches and Autonomous Timing Advance
The Unified Frame Structure concept supports various types of traffic and devices. For sporadic small
packet transmission, our vision is to use open-loop synchronization for battery and signaling
overhead saving reasons. Available downlink synchronization sequences and/or pilots are used to
establish a coarse downlink-based synchronization. Remaining timing offsets are resulting from the
unknown propagation round-trip times. The idea of Autonomous Timing Advance (ATA) [SW14] is
that the device uses estimates of its propagation delay (e.g. based on receive power levels of pilot
signals and/or some additional broadcast information by the base station for transmit power levels
or cell sizes) and coarsely pre-compensates its round trip delay by adapting its transmission timing
autonomously. This allows skipping closed-loop timing adjustments, which eases direct (e.g.
contention-based) data transmission via random access.
Two variants of ATA are possible:

Type 1: On the one hand the single devices may estimate the round trip time their signals are
suffering and apply a respective timing advance. When doing so, performance of the
respective transmission is improved.
Type 2: Another simpler variant is for all type II/III/IV devices to apply the same timing
advance, e.g. to half of the highest round trip time (or alternatively the 95%-tile of all existing
round trip times) appearing within the cell. When using this technique, inter-traffic
interference at the borders is reduced (in Fig. 3.3.1 results for the second variant are
presented). Various means are available for the devices in order to get knowledge on highest
round trip time. The most simple one is for the basestation to broadcast the value based on
measurements (e.g. during closed-loop synchronization of type I devices).
3.3.1 Relaxed Synchronization with UFMC

Figure 3.3.1 shows simulation results for an FDMA multi-user scenario. In the absence of CFO and
guard subcarriers, UFMC supports 33%-60% larger cell radii than CP-OFDM for the same MSE. When
using 2 guard subcarriers, the supported cell radii by UFMC are more than double the ones from
OFDM. With a CFO of 10%, OFDM cannot support the target MSE of -30 dB.
FP7 ICT GA 318555
D3.3
OFDM vs. UFMC, no CFO, w/ type 2 ATA applying N =0.5N OFDM vs. UFMC, 10% CFO, w/ type 2 ATA applying N =0.5N
TA OL,max TA OL,max
-20 -20 OFDM, LCP=79
OFDM, L =79
CP L = 80, no Guards
L = 80, no Guards L = 80, 1 Subc. Guard
L = 80, 1 Subc. Guard L = 80, 2 Subc. Guards
L = 80, 2 Subc. Guards L = 80, 5 Subc. Guards
L = 80, 5 Subc. Guards
-25 -25 dark gray:
] [dB]
dark gray:
] [dB]
UFMC w/ RC shaped
UFMC w/ RC shaped
windowing
windowing
inter
inter
light gray:
light gray:
E[MSE
UFMC w/o windowing
E[MSE
UFMC w/o windowing
-30 -30
-35
0 0.05 0.1 0.15 0.2 -35
0 0.05 0.1 0.15 0.2
N /N
OL,max N /N
OL,max
Figure 3.3.1 CP-OFDM vs UFMC using ATA type 2: Average distortion from FDMA multiuser interference as a
function of cell radius normalized to multi-carrier symbol duration. Left: no CFO, Right: CFO 10% of subcarrier
spacing. The dotted line represents a targeted maximum distortion of -30dB, determining the highest
supported normalized delay.
3.3.2 Relaxed Synchronization with FBMC

A flexible architecture for multiuser asynchronous reception on fragmented spectrum is able to
exploit the advantages of FBMC if the signal is efficiently demodulated in the frequency domain
without a priori knowledge of the FFT timing alignment (i.e. the location of the FFT block, this
property is called asynchronous FFT). A receiver architecture based on this assumption is depicted in
Figure 3.3.1. An asynchronous FFT of size KN is processed every blocks of N/2 samples generating KN
points, i.e. if rm is the mth received vector, a KN-point FFT is computed for samples k = (n + m N/2)
with n = 0, 1, . . . ,NK 1. These successive KN points are stored in a memory unit.
The detection of a start of burst is then achieved on the frequency domain (i.e. at the output of the
FFT) using a priori information from the preamble. CFO is first estimated using the pilot subcarrier
information of the preamble by computing the phase of the product between two consecutive FBMC
symbols at the location of the pilot subcarriers. The propagation channel is assumed static for the
duration of the burst. When large CFO correction is required, a first step in the estimation process
consists of scanning the subcarriers around the pilot subcarrier locations to determine the subcarrier
with the highest energy. A tracking algorithm of the CFO may complete the synchronization process
when the duration of the burst is large and the accuracy of the preamble based detection algorithm
does not meet the required level. CFO compensation is then performed in the frequency domain
using a feed-forward approach.
The channel coefficients are then estimated on the pilot subcarriers before being interpolated on
every active subcarrier. The use of a KN-point FFT makes the interpolation particularly specific to this
receiver. The estimation of the channel coefficients is performed before applying the FBMC
prototype filtering. The least square estimates of the channel coefficients are computed by applying
a CFO phase correction on the received signal according to the location of the pilot symbol on the
prototype filter. Performance may be further improved by considering all the positions of the
prototype filter and by applying a maximum ratio combining algorithm after interpolation.
Once the channel is estimated on all the active subcarriers, a one-tap per subcarrier equalizer is
applied before filtering by the FBMC prototype filter (section 3.4). Demapping and Log-Likelihood
FP7 ICT GA 318555
D3.3
Ratio (LLR) computation complete the inner receiver architecture. A soft-input Forward Error
Correction (FEC) decoder recovers finally the original message.
The asynchronous frequency domain processing of the receiver combined with the high stop-band
attenuation of the FBMC prototype filter provides a receiver architecture that allows for multiuser
asynchronous reception. FFT and Memory Unit are common modules, while the remaining of the
receiver should be duplicated as many times as the number of parallel asynchronous users the
system may tolerate.
Figure 3.3.2: Block diagram of the proposed multiuser asynchronous receiver architecture
3.4 D-PRACH
3.4.1 5GNOW one shot transmission concept

The IoT is expected to foster the development of 5G wireless networks and requires efficient access
of sporadic traffic generating devices. Such devices are most of the time inactive but regularly access
the Internet for minor/incremental updates with no human interaction, e.g. machine-type-
communication (MTC). Sporadic traffic will dramatically increase in the 5G market and, obviously,
such traffic should not be forced to be integrated into the bulky synchronization procedure of current
4G cellular systems [WJK+14]. Consequently, MTC traffic would be removed from standard uplink
data pipes with drastically reduced signalling overhead. This would improve operational capabilities
FP7 ICT GA 318555
D3.3
and network performance as well as user experience and life time of autonomous MTC nodes. The
new approach in 5GNOW is to use an extended PRACH which achieves device acquisition and
(possibly small) payload transmission "in one shot". Similar to the implementation in UMTS, the goal
is to transmit small user data packets using the PRACH, without maintaining a continuous
connection. So far, this is not possible in LTE, where data is only carried using the physical uplink
shared channel (PUSCH) so that the resulting control signalling effort renders scalable sporadic traffic
(e.g., several hundred nodes in the cell) infeasible. By contrast, in our design a data section is
introduced between synchronous PUSCH and standard PRACH, called D-PRACH (Data PRACH)
supporting asynchronous data transmission. Clearly, by doing so, sporadic traffic is removed from
standard uplink data pipes resulting in drastically reduced signaling overhead as well as reduced
complexity and power consumption of devices. The following approach is described in detail in
[WJK+14, KWJ+14, WKJ15, WGA15, WKJ+15].
We assume that each D-PRACH's data resource contains only a very few number of subcarriers like 5-
20 subcarriers. In addition, in a 5G system, we can expect that there is a massive number of MTC
devices which will concurrently employ these data resources in an uncoordinated fashion. In the
simplest approach the D-PRACH uses the guard bands between PRACH and PUSCH which is the focus
here. In an extended setting this region can be enlarged (by higher layer parameters) but, clearly, at
some point new efficient channel estimation must be devised different to the proposal in this paper.
Notably, in parallel work, we have set up a sparse signal processing approach to cope with this
situation [WJW14]. We show that waveform design in such a setting is necessary since the OFDM
waveform used in LTE cannot handle the highly asynchronous access of different devices with
possible negative delays or delays beyond the cyclic prefix (CP). Clearly, guards could be introduced
between the individual (small) data sections and to the PUSCH which, though, makes the approach
again very inefficient. Our results indeed show that up to six subcarriers can be obtained compared
to standard 4G OFDM setting.
To illustrate the approach, we consider a simple uplink model of a single cell network, where each
mobile station and the base station are equipped with a single antenna. We assume that there exist
the two channels PUSCH and the PRACH. On the PUSCH, the data bearing signals are transmitted
from synchronized users to the base station using Single Carrier Frequency Division Multiple Access
(SC-FDMA). A small part of the resources is reserved for PRACH, in which, at the first step of the
RACH procedure, users send preambles that contain unique signatures. The percentage of resources
for PRACH depends on the overall bandwidth, i.e., since PRACH always uses six physical resource
blocks (PRBs), e.g., for 1.4 MHz it occupies 100% of the particular sub-frame carrying PRACH, and for
20MHz it occupies 6% of the resources of the particular sub-frame. Our goal is not only to design
suitable waveforms for asynchronous PRACH operations, but also to leave PUSCH operations as
unaffected as possible. The general frame structure that we consider, more precisely, the time-
frequency resource grid for the described channels is illustrated schematically in Figure 3.4.1. To
minimize the interference between the channels, several subcarriers on both sides of the PRACH are
usually left zeros as a guard band. Moreover, we will exploit the PRACH to carry some data on the
guard bands. These users, however, may be completely asynchronous which can be a serious
challenge.
FP7 ICT GA 318555
D3.3
Figure 3.4.1: PRACH (blue) and PUSCH (red) regions. A guard interval (GI) separates PUSCH from PRACH in LTE
(gray). A part of this area is used to support data transmission of asynchronous users (green) in a novel D-
PRACH. The D-PRACH size can be variably determined by MAC
While we assume that the PUSCH signals are standard OFDM, we adopt for the PRACH signal the
BFDM approach which is well suited to sporadic traffic, since the PRACH symbols are relatively long
so that i.) the transmission is very robust to (even negative) time offsets and, b.) the side effects such
as spectral regrowth due to periodic setting when calculating the bi-orthogonal pulses are negligible.
In addition, BFDM is also more robust to frequency offsets in the transmission which, as well-known,
typically sets a limit to the symbol duration in OFDM transmission. Finally, the concatenation of
BFDM and several LTE PUSCH OFDM symbols together requires a good tail behaviour of the transmit
pulse in order to keep the distortion to the payload carrying subcarriers in PUSCH small. Conversely,
the dual pulse which accounts for the distortion of PUSCH onto PRACH can be controlled by iterative
interference cancellation (if necessary). This alleviates the typical problem of controlling
time/spectral localization of pulse and dual pulse. The excellent and controllable trade-off between
performance degradation due to time and frequency offsets is the main advantage of BFDM.
FP7 ICT GA 318555
D3.3
4 Performance Results
4.1 Waveform Performance and Parameter Optimization
4.1.1 Generalized frequency division multiplexing

New scenarios are being foreseen for 5G networks with requirements that cannot be addressed only
with throughput increment. GFDM is a promising candidate for the 5G PHY layer, because with its
flexibility the scheme can address the requirements of different scenarios. In the following, three
particular aspects are covered. First, the focus is laid on the compatibility of GFDM with the LTE
master clock of 30.72 MHz, because this feature will allow a smooth transition between 4G and 5G
by facilitating the reuse of existing hardware components. Further, MIMO is an important technique
which will be essential for any kind of 5G networks. However, potential self-interference is
prohibitive for a direct application of existing techniques. Hence, it is shown how time-reversal
space-time coding can be adapted for GFDM. Lastly, OQAM is a prominent extension of state-of-the-
art multicarrier waveforms like OFDM or FBMC. It will be shown how it can be also applied to GFDM.
4.1.1.1 GFDM parameterizations compatible with the clock rate of LTE

The evolution of mobile communication systems must take into account a soft transition between
generations. The LTE system uses a master clock frequency that is 8 times higher than the one used
in 3G, which has greatly simplified the development of hybrid devices, such as for 3G and 4G
[UUE+12]. It would be equally advantageous for manufacturers and telecommunication operators if
the 5G standard would also be based on the LTE clock rate. In this section we show that GFDM
signals can be integrated in the LTE time-frequency grid using the same clock rate of 30.72 MHz.
Moreover, low latency is an important feature for future wireless communication systems and GFDM
can be designed to reduce the impact of the PHY layer in the overall system latency. The main
approach is to reduce the GFDM symbol duration and add only a single CP for M subsymbols. The
latter property is particularly important, because prefixing each subsymbol would actually increase
the latency and reduce the spectral efficiency. The short subsymbol duration will lead to wider
subcarriers that might suffer from frequency selective effects of the mobile multipath channel. But,
because GFDM subcarriers have M times more samples in the frequency domain than LTE
subcarriers, it is still possible to achieve acceptable symbol error rate performance. In this section we
present two possible settings for GFDM to address two different scenarios. In the first one, there is
only a GFDM signal which is configured to address low latency, using the master clock of LTE and in
alignment with the LTE grid. The second approach allows the co-existence of a GFDM signal with
current LTE systems.
To achieve a low latency, it is advantageous if each GFDM symbol can be demodulated

independently. Hence, the channel estimation cannot rely on the system frame structure and each
symbol must carry all necessary information for single shot synchronization [GMM+14] and channel
estimation. In consequence, similar or even more rigid pilot placements are required for low latency
GFDM compared to the current LTE system. Nevertheless, throughput is not the main concern for
time-critical applications and the increased overhead due a large amount of pilot subcarriers per
GFDM symbol is acceptable in this scenario.
FP7 ICT GA 318555
D3.3
Although several improvements must be done also in the upper layers to achieve low latency,
reducing the transmission time interval is a good step towards lower system latency and GFDM is a
waveform flexible enough to address this aspect, as shown in the next subsections.
The parametrization presented here considers a 20 MHz LTE system operating on frequency division
duplex (FDD) mode as a reference. Table 4.1.1 shows the main parameters of the LTE system
[RSH11]. Other LTE signal bandwidths can be obtained by adjusting the number of active subcarriers.
Table 4.1.1: LTE parameters for the FDD mode
The LTE time-frequency grid is organized in resource blocks (RBs), each one having a bandwidth of
180 kHz (12 subcarriers) and a duration of 0.5 ms (7 or 6 OFDM symbols for the normal and extended
modes, respectively). A resource block is the basic quantity for resource allocation in LTE, which
should be respected by a future 5G system. Figure 4.1.1 depicts the LTE resource block structure
assuming normal operation mode.
Figure 4.1.1: Time-frequency structure of the LTE resource block on normal operation mode
Now, the main objective is to integrate the GFDM signal in the time-frequency grid of the LTE system
while enabling low latency applications. The duration of the GFDM symbol must be reduced to an
integer fraction of the 1 ms subframe duration and a set of GFDM subcarriers must fit in an integer
multiple of 180 kHz. Table 4.1.2 presents one possible set of GFDM parameters for this scenario.
FP7 ICT GA 318555
D3.3
Table 4.1.2: GFDM configuration aligned with the LTE grid.
From Tables 4.1.1 and 4.1.2, it can be seen that the proposed GFDM approach has the same
subframe duration as the LTE grid. Nevertheless, the GFDM symbol duration is 7.5 times smaller than
the equivalent slot duration of the LTE system. Assuming that the GFDM system is designed for
independent demodulation of each symbol, it is possible to reduce the contribution of the PHY layer
modulation to the latency by a factor of 15, compared with current LTE frame structure (LTE requires
at least two slots of 0.5 ms to demodulate the user data). Of course, the total latency of the system
also depends on a number of factors that cannot be addressed in this section. However, Table 4.1.2
clearly shows that the GFDM PHY can contribute to the latency reduction while keeping the same
time-frequency grid as the 4G system. Notice that three GFDM subcarriers occupy the bandwidth of
four LTE resource blocks. This means that each GFDM subcarrier is 16 times wider than the LTE
subcarriers. Because each GFDM subcarrier has M=15 times more samples than an LTE subcarrier,
the spectrum resolution of both systems is approximately the same. GFDM employs a slightly smaller
CP length than LTE and it does not require a larger CP for the first GFDM symbol. The Dirichlet pulse
makes the system orthogonal and, because the roll-off factor in the frequency domain is zero, the
GFDM subcarriers do not overlap with the surrounding subcarriers outside of the used resource
blocks. Therefore, the LTE time-frequency grid can be used to accommodate GFDM signaling as
shown in Figure 4.1.2. Note that the conventional GFDM subcarrier spacing N=K has been used here
to match the LTE frame structure.
Figure 4.1.2: GFDM frequency grid to match the LTE frame structure
FP7 ICT GA 318555
D3.3
GFDM can also be configured as a secondary system in the LTE grid, i.e. to use two empty resource
blocks, leaving a guard band to avoid interference in the surrounding resource blocks that are used
to transmit the conventional LTE signal. In this case, the GFDM signal can be seen as a secondary
signal that is used to explore vacant resource blocks for low latency applications. Table 4.1.3 shows
the GFDM parameters for this approach.
Table 4.1.3: Parameters for asynchronous GFDM signaling
A new approach to generate the GFDM signal must be introduced here in order to keep the
subcarrier spacing compatible with the LTE time-frequency grid. The subcarrier bandwidth is 320 kHz,
while the subcarrier spacing must be a multiple of 180 kHz (bandwidth of one RB). In order to
achieve this frequency spacing, N must assume a value that differs from subsymbol spacing K.
Nevertheless, N must be carefully chosen in order to guarantee that an integer number of subcarrier
cycles is present within the duration of one GFDM frame, otherwise there will be phase jumps
between the CP and the GFDM signal, leading to a strong out-of-band emission. The parametrization
presented in Table 4.1.3 achieves this goal.
From Tables 4.1.1 and 4.1.3 it can be seen that the GFDM subcarriers are 21.33 times larger than LTE
subcarriers in terms of bandwidth, while the GFDM symbol duration is 10 times smaller than the
corresponding slot duration of the LTE system. Moreover, GFDM subcarrier resolution in the
frequency domain is M=15 times the resolution of the LTE subcarriers. The CP length has been
shortened to 2/3 of the LTE CP length, which means that this approach is appropriate for small cell
size (typically for diameter smaller than 4 km).
LTE equipment transmits system information periodically and even on empty resource blocks, which
makes it difficult to use the approach proposed in Table 4.1.3. One solution is to consider a LTE
system operating in a spectrum hole that is larger than the system bandwidth, e.g., a 5 MHz signal
being transmitted in the center of a 10 MHz band. Since the LTE grid is the same for any bandwidth
configuration, with the only difference being the number of active subcarriers, the GFDM signal
presented in Table 4.1.3 can be appended on the edges of the LTE signal, as depicted in Figure 4.1.3.
FP7 ICT GA 318555
D3.3
Figure 4.1.3: GFDM as a secondary signal in the LTE time-frequency grid
Notice that the low out-of-band (OOB) emission of GFDM causes small interference in the LTE signal.
However, the high OOB emissions of the LTE OFDM signal might be harmful for the GFDM signal
[MLR+11]. The interaction of the LTE OOB with GFDM signals must be considered in order to specify
the forward error control codes and other protective measures for the GFDM PHY layer.
4.1.1.2 Multi-User MIMO-GFDM

The parameters used to evaluate the symbol error rate (SER) performance of the system described in
Section 2.2.3 are presented in Table 4.1.1.2.1. The power delay profile of the wireless channel is
shown in Table 4.1.1.2.2, which is based on the Extended Pedestrian A (EPA) model [RSH11]. The
channel taps are multiplied by i.i.d. complex Gaussian variables with variance 1 , creating
independent Rayleigh fading for each user and antenna. The channel remains constant during the
transmission of two GFDM blocks. All users are synchronized with the base station's master clock, i.e.
no frequency misalignment occurs. The timing misalignment between users and base station is
uniformly distributed between [-3,3] samples. Notice that the CP and CS are large enough to
accommodate the channel impulse response and the time misalignment.
Table 4.1.4: Simulation parameters
Table 4.1.5: Channel power delay profile in the simulations
FP7 ICT GA 318555
D3.3
An upper bound of the TR-STC-GFDM SER performance under a frequency-selective fading channel is
given by [MLG+15]
i
JL 1
JL 1 i 1
pe 4 , (4.1.1.2.1)
i0 i 2
where
3
2
e
Es
JL
J 11 J 1 0N 0
, ,
3
2
J 2 2 e
Es
J 1 0N 0
J is the modulation cardinality, E S is the average symbol energy, N 0 is the noise spectrum density,
0 is noise enhancement factor (NEF) , J is the number of transmit, L is the number of receive
antennas, and

2 2
e
E [| h n | ].
n
When the different amount of CP in GFDM and OFDM is considered in the calculation of 0 , we
obtain
K N CP N CS K M N CP N CS
0 ,O F D M K
0 ,G F D M KM
,
which shows that, for certain configurations of K and M and depending on the additional factor
A A
1 1
H
1
tr that is a specific parameter of the GFDM subcarrier filter, GFDM can reach
N
a better spectral efficiency compared to OFDM. Eq. (4.1.1.2.1) has been adapted from STC-OFDM SER
performance [MB10] by adding the NEF and it becomes a tighter upper bound if J L 2 and the
channel frequency response is flat per subcarrier. For the SISO case ( J L 1 ) the gap between the
SER and the upper bound increases with the modulation order.
Fig. 4.1.4 compares the SER performance of conventional STC-OFDM [LW00] and TR-STC-GFDMA
with perfect CSI at the BS. It shows that TR-STC-GFDMA users with two transmit antennas
(users 4 to 8) achieve the same diversity gain as STC-OFDM, however, due to the more efficient use
of the CP, the SER curve of GFDM is shifted 0.25 dB to the left. As expected, users with one transmit
antenna (users 1 to 3) do not benefit from transmit diversity gain.
Fig. 4.1.5 shows the SER performance with the proposed channel estimation algorithm presented in
equation (2.2.3.2.1) of Section 2.2.3.2. When a 3 dB pilot boost is used, the performance is
approximately 1.8 dB worse compared to perfect CSI but no further performance degradation in
diversity is apparent.
FP7 ICT GA 318555
D3.3
Figure 4.1.4: SER performance of STC-OFDM and TR-STC-GFDMA. No error correction coding was applied
Figure 4.1.5: SER performance of TR-STC-GFDMA considering the channel estimation
4.1.1.3 Adapting GFDM for time-shifted OQAM signalling

Offset QAM (OQAM) modulation is a powerful extension that can be applied to state-of-the art
modulation schemes like OFDM or FBMC. It allows combining excellent spectral properties with the
elegance of orthogonal signalling. In the following, we show how the GFDM modulator and
demodulator can be adapted to support offset modulation.
FP7 ICT GA 318555
D3.3
In time-shifted OQAM-GFDM, the complex-valued data symbols d k , m i k , m jq k , m are transmitted

on K subcarriers and M subsymbols, where real and imaginary part are offset by K
2
samples. Each
subsymbol is pulse-shaped with a prototype filter g[n] [Mat15] and the transmit signal is given by
M 1 K 1 M 1 K 1
( k m od 2 ) ( k m od 2 )
x[ n ] j ik , m g k , m [ n ] j j q k ,m g
k ,m
1
[ n ], (4.1.1.3.1)
m 0 k 0 m 0 k 0 2
where N=MK, and
g k ,m [ n ] g ( n m K ) m o d N w
kn
,
2
is the pulse shape of the kth subcarrier and mth subsymbol with w e x p ( j ) . The parameter
K
{ 0 , 1} assumes 1 for symmetric real-valued pulse shape, or 0 for non-symmetric conjugate root
pulse shape [Dem98,TB04].
At the receiver, a matched filtering operation is carried out, i.e.
( k m od 2 )
ik , m
kt
( x[ n ] j # g [ N n ] ) |n m K
*
w
( k m od 2 ) kt
( x[ n ] j # g [ N n ]) |
*
q k , m w 1
,
n m K
2
where # is the circular convolution.
The projection between the kth and (k'+k)th subcarrier
X [m , k ] g[n] j k
w
kn
* g [ N n ] |n m K
*
can achieve the real-orthogonality conditions
X [ m , k ] j X [ m , k ] [ k , m ]
1 1
jX m , k X m , k 0

2 2
at all k,m for half-Nyquist filters [FAB95], which shows that it is possible to transmit an interference-
free signal with a time-shift of half subsymbol.
The time-shift OQAM-GFDM can also be represented by matrix operations. In this case, the transmit
vector equivalent to equation (4.1.1.3.1) is given by
x A J { d } jB J { d } , (4.1.1.3.2)
where d contains the complex-valued data symbols to be transmitted. The auxiliary matrix
k
J IM d ia g ( j ) denotes the equivalent phase shift among subcarriers, where d i a g ( a ) returns a
FP7 ICT GA 318555
D3.3
matrix with a on its main diagonal and zeros otherwise, is the Kronecker product and I M is an M
size identity matrix. The columns of the matrix A carry all possible pulse-shapes g k , m [ n ] .
Equivalently, the columns of B are composed by g k , m 1 / 2 [ n ] . Notice that the first term of (4.1.1.3.2)
corresponds to the first double sum in (4.1.1.3.1), while the second term in (4.1.1.3.2) corresponds to
the second double sum in (4.1.1.3.1).
Assuming a perfect channel, the received data symbols at the output of the matched filter receiver
are given by
d {(A J ) y } j { ( B J )
H H
y} .
Clearly, this approach leads to an implementation where the time-shift OQAM-GFDM can be seen as
two independent classic GFDM transceivers, as depicted in Fig. 4.1.6.
Figure 4.1.6: Block structure of the time-shift or frequency-shift OQAM-GFDM.
4.1.2 UFMC
4.1.2.1 Parameter sets

UFMC can be adjusted to completely fit into the LTE(-A) signal structure, where a complete reuse of
all existing pilot patterns, control signaling resource elements, precoding, encoding/decoding
including rate matching etc. is possible.
As LTE parameters w.r.t. subcarrier spacing and cyclic prefix length are properly designed for macro
cells, they already provide a reasonable choice for UFMC as well. (On top, UFMC allows for further
per-subband adaptation.)
Table 4.1.6 UFMC parameters fully adjusted to LTE grid for 20/10 MHz
Subframe duration 1 ms or 30720/15360 samples
UFMC symbol duration 2048/1024 + L -1 samples
Subcarrier spacing and bandwidth 15 kHz
Sampling frequency 30.72/15.36 MHz
# symbols per subframe 14
CP length None
Filter length L = 161/81 for symbol #1 and #8, L=145/72 for other
symbols
[Option: L = 147/74 for all symbols and 4/2 zero
samples at beginning of subframe]
Subband filter Dolph-Chebychev, 40 dB side lobe attenuation
FP7 ICT GA 318555
D3.3
UFMC Filter Optimization for certain expected ranges of carrier frequency offsets (e.g. resulting from
oscillator requirements) is described in section 4.1.2.
4.1.2.2 Time Frequency Efficiency

E.g. with N = 1024, L=74, UFMC operates with 10 MHz LTE-like parameters on the same symbol rate
and subcarrier spacing as CP-OFDM in LTE, but being able to allocate roughly 10% more
subcarriers than CP-OFDM [SWC14]. This is demonstrated by Fig. 4.1.6. Due to the long filter length,
FBMC has a time-overhead dependency as a function of the burst length which is not the case for CP-
OFDM and UFMC. This means, UFMC, like CP-OFDM, allows for the efficient insertion of small control
elements, occupying one or a few resource elements, like sounding symbols, control information
(e.g. ACK/NACK feedback) etc.
Ng = 15, SLA = 60 dB, N=1024

0.95
0.9
0.85
0.8
0.75
TF
UFMC, L = 40
r
0.7
UFMC, L = 60
0.65 UFMC, L = 80
FBMC, K = 4
0.6 FBMC truncated by 0.5N
FBMC truncated by 1.5N
0.55 FBMC truncated by 2.5N
CP-OFDM, LTE settings
0.5
2 4 6 8 10 12 14 16 18 20
M
Figure 4.1.6- Time-frequency efficiency of FBMC, CP-OFDM and UFMC as a function of burst length
M (in number of multi-carrier symbols): Number of usable resource elements per unit time-frequency bin
We now compare the efficiency of UFMC and OFDM in a fragmented spectrum scenario under
different levels of fragmentation. The obtained efficiency is based on the fraction of usable
subcarriers w.r.t. the base band signal same radiated power at the band edge. The number of
required guard carriers are obtained using Table I in [SWC14] with L=80, SLA = 40 dB from a 10 MHz
setting. This assumes that any subsequent RF-(pre)processing, including clipping, filtering plus any PA
non-linearities have the same impact for all waveforms when it comes to fulfilling spectral mask
requirements for identical radiated band edge power.
Note that the criterion identical radiated band edge power does not take into account the impact
for the RF pre-processing filter slope, so we consider this result, depicted by Fig. 4.1.7 as a rough
approximation.
FP7 ICT GA 318555
D3.3
Efficiency in fragmented spectrum

1
Efficiency: Fraction of usable subcarriers

0.9
0.8
0.7 OFDM
UFMC
0.6
0.5
0.4
0 5 10 15 20 25 30 35 40
Fragment size (in MHz)
Figure 4.1.7 Frequency-efficiency (fraction of usable subcarriers) of CP-OFDM and UFMC as a function of
spectrum fragment size
4.1.2.3 UL CoMP Multiuser Joint Reception under Carrier Frequency Offsets

Results for multi-user uplink distributed MIMO, thus coordinated multi-point (CoMP) joint reception
were provided by [D3.1] and [VWS+13]. They showed a higher robustness of the UFMC waveform
compared to CP-OFDM in the presence of carrier frequency offsets.
-1
10
-2
10
QPSK SER
UFMC no CFO
-3
10 UFMC perf. CFO comp
UFMC CFO comp w/ err
OFDM no CFO
OFDM perf. CFO comp.
-4 OFDM CFO comp. w/ err
10
12 14 16 18 20 22 24 26
Eb/N0
Figure 4.1.8: SER vs Eb/N0 for a 2 user / 2 cell CoMP scenario: Comparison between UFMC and OFDM for
absence of carrier frequency offset (no CFO), perfectly known and compensated carrier frequency (per. CFO
comp) and a CFO compensation with a residual mismatch of 10% of the total CFO.
Figure 4.1.8 shows the performance w.r.t. QPSK symbol error rate over Eb/N0 for a 2 user / 2 cell
uplink joint reception scenario using a linear zero forcing receiver across all antennas and subcarriers.
For details of the simulation setting please refer to [D3.1].
FP7 ICT GA 318555
D3.3
4.1.2.4 Filter Optimization

UFMC results in previous 5GNOW deliverables have been based on Dolph-Chebychev filters. Those
filters are parametrizable in terms of filter length and sidelobe level attenuation. The side lobe level
attenuation is in fact trading off the main lobe width against the side lobe level of the filter. Dolph-
Chebychev filters are a reasonable approach: For a given side lobe level, the main lobe width is
minimized.
In D3.3 we now aim at optimizing the filter coefficients for UFMC for certain design targets, not
restricting ourselves to Dolph-Chebychev. The main target in our solution [WWS+14] is to increase
the robustness against inter-carrier interference from carrier frequency offsets (CFO). This is helpful
in the 5GNOW framework of relaxed synchronicity, provides better support for low end devices (e.g.
MTC, allowing for cheaper oscillators) and supporting shorter wake-up times in sporadic
transmission, as oscillators tend to be less accurate after wake-up.
The considered inter-carrier interference (ICI) is in a multi-user uplink FDMA scenario, using the
UFMC waveforms. Each device has individual CFOs, destroying the subcarrier orthogonality between
different user allocations. The filter optimization aims at mitigating ICI impairments.
The usage of this filter optimization in a practical 5G system is intended to be computed offline. A
limited set of UFMC waveforms should be used, addressed e.g. by a codebook of 2-3 bits, thus
providing 4-8 different parameter sets. Those parameter sets now should be optimized for a certain
scenario and set of requirements.
In this section we address two different optimization criteria:

1. Maximizing signal-over-out-of-subband-leakage ratio (SLR)
2. Maximizing signal-over-inband-distortion-plus-out-of-subband-leakage ratio (SDLR)
The optimization approach is in detail described in [WWS+14]. It leads to a Generalized Eigenvalue
problem which can be computed in a closed-form solution. The SLR criterion has shown to provide
the better performance.
In Figure 4.1.9 we depict the SLR optimization outcome for a deterministic relative carrier frequency
offset (rCFO). We observe that the optimal filter shape is similar for a broad range of rCFOs. The
optimization outcome is similar to a particular choice of a Dolph-Chebychev filter here with 13.4 dB
of side-lobe level attenuation.
Time Domain Frequency Domain

0.6 20
rCFO0.1
rCFO-0.1 0
0.5
rCFO0.2
rCFO-0.2
rCFO0.3 -20
Magnitude/dB
0.4 rCFO-0.3
Amplitude
rCFO0.4
rCFO-0.4 -40
rCFO0.5
0.3 rCFO-0.5
-60
0.2 optimized
-80
chebwin(16,40)
chebwin(16,13.4)
0.1 -100
Figure 4.1.9 SLR optimized filter for deterministic rCFO normalized to subcarrier spacing
FP7 ICT GA 318555
D3.3
In Figure 4.1.10 we have carried out the SLR optimization for an entire range of rCFOs from
[-0.5,,+0.5]. This approach thus can e.g. suit a certain oscillator requirement range coming from a
future 5G system. Certainly, the filters are not optimized for a certain deterministic rCFO, but should
cope with a whole range of rCFOs occurring in the system.
Time Domain Frequency Domain

0.6 20
optimized optimized
chebwin(16,12.1) 10 chebwin(16,12.1)
0.5
0
Magnitude/dB
-10
0.4
Amplitude
-20
0.3
-30
-40
0.2
-50
0.1 -60
Figure 4.1.10 SLR optimized filter for uniformly distributed rCFO between 0.5 subcarrier
Spacing
Again, a similarity to Dolph-Chebychev filters (here with side-lobe attenuation of 12.1dB) can be
observed.
Let us now look at the resulting performance gain of the filter optimization for two different
scenarios:
Scenario 1: The receiver does not account for CFO - not recommended and just provided as
reference.
Scenario 2: The receiver estimates the CFO and runs an appropriate phase de-rotation to
compensate for it. This should not pose a problem to a 5G base station.
Note that in Scenario 2 is still residual inter-carrier interference in a multi-user FDMA scenario, as the
subcarriers of different user allocations with different rCFOs are not orthogonal anymore. For this
purpose the filter optimization aims at mitigating the resulting ICI by maximizing SLR. The resulting
relative performance gains of UFMC over CP-OFDM are shown in Figure 4.1.11. When using the SLR
criterion, around 10 dB of SIR improvement can be achieved over CP-OFDM with typical rCFOs of up
to 0.1. The optimized SLR solution still is more than 5dB better in terms of SIR, compared to a Dolph-
Chebychev filter with 40dB side lobe attenuation.
FP7 ICT GA 318555
D3.3
Scenario 1 Scenario 2
1 12
10
0.8
8
SIR Gain/dB
SIR Gain/dB
0.6
6
0.4
4
0.2 chebwin(16,40) 2 chebwin(16,40)

SDLR SDLR
SLR SLR
0 0
0 0.2 0.4 0 0.2 0.4
rCFO rCFO
Figure 4.1.11 SIR gain of UFMC over CP-OFDM as a function of CFO normalized to subcarrier spacing for
receiver without CFO compensation (Scenario 1) and receiver with CFO compensation (Scenario 2). Optimized
filters, using SDLR/SLR criteria, are compared to Dolph-Chebychev filters with 40dB side lobe level attenuation
(denoted chebwin(16,40))
To sum up, filter optimization can help to tailor UFMC to the respective requirements and further
improve its performance advantages over CP-OFDM.
4.1.3 FBMC
The FBMC PHY parameters are derived from the LTE format 10MHz bandwidth with sampling rate
15.36MHz. Details on the selected parameters are given on the table below.
Simulation Parameter Value

Subcarrier spacing 15 kHz LTE format for
Subcarriers per PRB 12 10MHz bandwidth
FFT size 1024 with sampling rate
No. of used subcarriers 12 to 501 15.36MHz
No. of used PRBs Configurable
Block (subframe) duration 1.6ms
FBMC filter K=4 (optimized for ACLR)
Modulation QPSK to 64-QAM Adaptive data rate
Coding rate 1/2, 2/3, 3/4 (CC K=7)
OFDM CP-length N/A
Performance metric (BER,PER) vs SNR (freq. and timing offsets)
Table 4.1.7: FBMC PHY parameters
FP7 ICT GA 318555
D3.3
4.1.3.1 Summary of previous results

In the first part of this section, we remind the main FBMC performance results that were described in
[D3.2]. These results apply both for downlink and uplink transmissions.
Particular attention was paid to CFO compensation in order to relax synchronism requirements
beyond one carrier spacing. It was shown that frequency synchronization (estimation and
compensation of the CFO) with FBMC can be entirely realized at the UE in the frequency domain. The
most part of the CFO can be easily and accurately estimated thanks to a simple energy detection
algorithm on the preamble carriers and thus mitigation of Inter Carrier Interference (ICI) is achieved
at low cost. The residual part of the CFO after this first coarse estimation is lower than 100/(2K)%
(12.5 % with the overlapping factor K equal to 4) of carrier spacing. The same method applied to
OFDM would lead to a much higher residual CFO of 50 % at maximum and thus worst performance in
terms of BER would be obtained. Furthermore, the algorithm proposed to estimate the residual part
exhibits reduced complexity while achieving good performance. Depending on the receiver target
SNR, complexity may be traded-off to keep Relative Mean Square Error (RMSE) introduced by CFO
below thermal noise.
The performance of the proposed algorithm is analysed in this section using the following FBMC
parameters: N=1024 and K=4. Figure 4.1.3.2 gives the architecture of the proposed implementation.
The first step, performed by the Shift module, consists in correcting the integer part q . The result is
then filtered by the 2 Q 1 -tap filter W . A phase correction performed on each FBMC symbol
completes the CFO correction process.
Figure 4.1.12: Block diagram of the proposed CFO correction scheme
The RMSE for values (normalized CFO) from 1% to 12% and for different values of Q is given in
Figure 4.1.3.2. W is derived using the truncated ZF method.
FP7 ICT GA 318555
D3.3
Figure 4.1.13: RMSE performance of the proposed filtering scheme
When Q = 0 , the algorithm only performs phase correction. The benefit of ICI mitigation is clearly
demonstrated. Parameter Q could be chosen as a function of the system Signal to Noise Ratio (SNR)
so as to limit the system by thermal noise but not by interference. In practice, the estimation of the
CFO is never perfect. As illustrated in Figure 4.1.13, a residual CFO of 1% generated by the correction
mismatch exhibits a RMSE power level below -30 dB and does not degrade performance significantly.
Channel interpolation was also carefully considered. As multiuser asynchronous FDMA generates
heavily fragmented spectrum blocks, channel estimation should be optimized for the edges of the
receiver active carrier bands. A performance improvement of up to 7dB on the RMSE of the
estimated channel may result on some simulated scenarios, as can be seen in Figure 4.1.14.
Figure 4.1.14: RMSE as a function of the channel delay spread for a SNR of 15 dB and for different
interpolation filter blocks
Finally, a new equalizer scheme was thoroughly presented. Its complexity is contained while good
performance for channel exhibiting large delay spread is achieved, as can be seen in Figure 4.1.15. As
FP7 ICT GA 318555
D3.3
a comparison, using the 10 MHz LTE parameters, the receiver performs well for channels with
maximum delay spread of up to 8.3 s (time arrival of taps is uniformly distributed and their complex
coefficients follow a Rayleigh distribution). This compares with standard LTE that been design for
channels with delay spread of up to 4.7 s. Moreover, we have demonstrated that the proposed
equalizer does not required FFT synchronization and is therefore adapted to asynchronous multiuser
reception.
Figure 4.1.15: Performance of the proposed receiver for various interpolation filters as a function of
channel delay spread
In the second part of this section, we remind the main FBMC performance results that were
described in [D3.1]. These results apply for downlink CoMP.
Joint transmission among multiple cells can provide significantly large throughput gains at the users.
Nevertheless [MP11] points out the price to pay to achieve an efficient CoMP transmission with
OFDM waveform. One of these identified drawbacks is the precise synchronization of the oscillators,
for e.g. related to the synchronous data exchange and clock synchronization between eNBs for
proper timing of coordinated transmission. On the other hand, the FBMC (Filter Bank Multi Carrier)
waveform with its spectrally well shaped prototype filters and overlapped time symbols has some
inherent features which make it a natural choice for some of the anticipated 5G application
scenarios. First of all it does not require a cyclic prefix and has an almost perfect separation of
frequency subbands without the need for strict synchronization. Consequently its properties make it
especially suited for fragmented spectrum and Coordinated Multi Point (CoMP)
Transmission/Reception. In [CDK13][D3.1] it is shown that both the Carrier Frequency Offset (CFO)
estimation and compensation can be realized at the UE side, in the frequency domain. These
operations require no feedback from the UE to the BSs. [CDK13][D3.1] also demonstrated that FBMC
allows an easier and more accurate estimation and compensation of the CFO than OFDM. One other
issue is the necessary feedback for time synchronization at the UEs. To do so and when dealing with
large cooperating areas, the size of the cyclic prefix can reach high values, leading to high spectral
efficiency loss for OFDM. In [CDK13][D3.1], FBMC modulation was demonstrated to be very resistant
to time propagation differences between signals from two Base Stations, due to its overlapping
FP7 ICT GA 318555
D3.3
structure. As can be seen in Figure 4.1.16, delays up to 120 samples, i.e. 7.8 s (2340 meters) can be
tolerated at the UE without any estimation nor correction.
Figure 4.1.16: BER of CoMP FBMC with MRC at the receiver for different delays between BSs
4.1.3.2 Capacity of multicarrier asynchronous uplink transmission

In the following, we present additional analytical and simulation results that were not described in
the previous deliverables, with a specific focus on capacity for uplink asynchronous transmissions.
The estimation of the channel capacity is a useful criterion for comparing the performance of digital
modulation schemes. The channel capacity of any M-ary Quadrature Amplitude Modulation (M-
QAM) scheme rises to a maximum of log 2 M bits per symbol as the SNR increases. The channel
capacity of M-QAM constellation in presence of Additive White Gaussian Noise (AWGN) is given by
[Ung82]:
CM QAM
( ) =
M
M 2 td ( ) d ( ) 2 (4.1.3.1)
M
2 2
e 2
dt
2 t mi mi
log log e
2
M m =1 i = 1
where d mi is the distance between each real (resp imaginary) point:
Re ( X ) Re ( X i )
d mi ( ) =
m
(4.1.3.2)

and where 2 is the noise variance and X i the QAM symbols.Assuming that interference generated
when timing misalignment occurs at the receiver side and thermal noise are independent, the
equivalent noise variance my be written as:
( , p ) = n i ( , p )
2 2 2
(4.1.3.3)
where n is the standard deviation of the thermal noise and i ( , p ) is the standard deviation of
the interference at carrier index p due to the misalignment of . It should be mentioned that the
thermal noise is independent of p or as AWGN is considered.
Under the assumption of synchronous transmission over N a active carriers, the capacity of a
multicarrier waveform is given by:
FP7 ICT GA 318555
D3.3
p=N 1
a
1
C MC ( n , ) = CM QAM
( ( , p )) (4.1.3.4)
N a p=0
Note that the above formula for the capacity also holds for multicarrier O-QAM systems: with such
systems, real and imaginary parts of the N a M-QAM symbols are transmitted on two adjacent
carriers, each one with half capacity.
The effect of asynchronous transmission is modeled on capacity computation by averaging the
capacity over all the possible timing misalignment. In cellular network, the time misalignment is the
consequence of the propagation delay between user equipments and the base station. If the users
are equidistributed in the cell a uniform distribution of can be considered. Therefore the capacity
is given by:
p= N 1
a
1 1
C async ( n ) =
max
( ( , p )) d (4.1.3.5)
max

0
N
CM QAM
a p=0
It should be mentioned that Eq. (4.1.3.5) is valid, if and only if interference terms are uncorrelated
with a probability density function following a Gaussian distribution. In practice interference terms
are correlated as interference is generated by a misalignment of the FFT window. However by
simulations we have checked that the assumption of uncorrelated noise is valid.
Scenario description
We have considered a scenario representing fragmented spectrum access in the context of
asynchronous uplink. A set of three users is simulated as illustrated in 4.1.17. We assumed perfect
synchronization and perfect channel estimation of the user of interest (User Equipment 0 UE0).
N a = 24 carriers (or 2 resource blocks) are allocated to UE0, and 120 carriers for UE1 and UE2. UEs are
considered unsynchronized. The intercarrier spacing is set to 15 kHz, N=1024 and for OFDM
waveform a guard interval (GI) of 72 samples (approximately 1/14 x N) is assumed. These parameters
have been inspired by possible extensions of LTE parameters. No guard carriers are considered for
scenario (a) while one guard carrier is inserted at the edges of each user spectrum bands for scenario
(b). In the following section, we compared the capacity in case of OFDM and FBMC waveform for the
two scenarios. The cell coverage is fixed and assumed to be of 5 km. In that case maximum timing
offset is equal to max = 16.67 us or 256 T s , where T s is the multicarrier sampling period
1
( s ).
1024 15 10
3
Figure 4.1.17: Proposed asynchronous scenario with three users (a) without guard band, (b) with
guard bands
FP7 ICT GA 318555
D3.3
Interference evaluation
In order to determine the capacity of the proposed configurations, we have first determined the level
of interference generated by neighboring UEs as a function of the timing offset. For the FBMC
waveform a phase factor had to be added to the interfering UE in order to break orthogonality at the
edges of the band. In that case, the interference term is computed by averaging a uniformly
distributed phase in the range [0,2 [ . The Signal to Noise Ratio (SNR) is set to 28 dB. We depicted in
Figure 4.1.18 the Signal to Noise plus Interference ratio (SINR) measured at the output of the
equalizer for OFDM and FBMC waveforms in the case of scenario (a), i.e. without guard carriers. The
SINR has been computed for each carrier location and for each timing offset from 0 to 256 T s , by
steps of T s under the assumption of a perfect power control (i.e, each user is received with the same
power).
Figure 4.1.18: SINR in dB as a function of carrier index and timing offset for OFDM and FBMC
waveforms assuming scenario (a).
For OFDM, thermal noise dominates the SINR when the timing offset is lower than the duration of
the guard interval. In that case, the received signal is not affected by interference. However, as soon
as the timing offset is larger than the guard interval, SINR is dominated by the interference. It clearly
illustrates that the interference level is more important on the edges of the spectrum when the
OFDM signal is not synchronized within the guard interval of its adjacent signal. This is due to the sinc
frequency shape of the OFDM waveform.
For FBMC, results are different. The center of the spectrum is always dominated by thermal noise,
while interference at the edges of the carrier is preponderant, whatever the timing offset. This is a
direct consequence of the properties of the prototype filter which has been designed to minimize
out-of-band interference. Interference affects the edges of the spectrum because the orthogonality
of OQAM modulation is not preserved between asynchronous adjacent users.
FP7 ICT GA 318555
D3.3
SINR (without thermal noise) on one RB was also evaluated when adding carrier frequency offset of
respectively 0, 5, 10 and 30%, i.e. respectively 0, 750, 1500 and 4500 Hz. The curves are plotted on
the figure below (see Figure 4.1.19), where the blue curve is for FBMC and the red curves are for
OFDM. The x axis is the normalized timing offset. What can be seen is that CFO does not have impact
on FBMC thanks to the applied guard carrier on the edge of the RB whereas for OFDM, the
performance are severly impaired within the guard interval when CFO is applied.
Figure 4.1.19: average SINR in dB versus timing offset and CFO, for OFDM and FBMC.
Capacity computation
Once the level of interference is evaluated, the capacity of the transmission may be derived. We
depicted in Figure 4.1.20 the capacity of OFDM and FBMC in the case of scenario (a) and (b), using
Eq. (4.1.3.5) for various SNR values and modulation orders. Perfect power control is here assumed. It
should be noted that we do not take into account in the capacity calculation the loss due to the
guard interval (OFDM) and the filter rising and falling time (FBMC). Capacity is equivalent when the
number N s of symbol is equal to:
( K 1) N
N s
= (4.1.3.6)
GI
where GI , is the duration of the guard interval expressed in number of samples. For the rest of the
analysis, we assumed that the loss due to the cyclic prefix is equivalent to the loss due to filter rising
and falling time.
In the case of QPSK modulation (asymptotic capacity of 2), without guard carrier, the capacity is close
to the one of synchronous transmission. In that case the level of interference is much lower than the
required SNR to allow the decoding of QPSK. For 16-QAM modulation (asymptotic capacity of 4),
FBMC gives a significantly better capacity, particularly for SNR values above 10 dB. Due to the better
frequency localization, only the carriers located at the border of the user spectrum are affected by
interference. These results also demonstrate that performance is limited by interference as the
capacity of synchronous transmission is never reached. However, in the proposed scenario, capacity
loss due to signaling has not been taken into consideration. The impact of this hypothesis should also
be taken into account in the capacity computation for a fairer comparison.
FP7 ICT GA 318555
D3.3
Figure 4.1.20: Capacity in bit/s/carrier for OFDM and FBMC waveforms assuming scenario (a) and (b).
Blue, red and green curves stand for respectively QPSK, 16-QAM and 64-QAM modulations.
For 64-QAM modulation (asymptotic capacity of 6), the FBMC waveform clearly outperforms the
OFDM waveform. Interference dominates the SINR for the OFDM waveform and consequently for a
given capacity of 5 bits/s/carrier, the SNR loss is of around 5 dB.
We also illustrated in Figure 4.1.20 the capacity comparison for scenario (b), i.e. with guard carriers.
The achievable capacity is evaluated assuming a constant power on the bandwidth. In other words,
the power of each carrier is boosted in the case of guard carriers. Moreover, the loss in capacity as a
consequence of the use of guard carrier is taken into account into the calculation. For the OFDM
case, as the power on each carrier is increased, the level of interference follows. Results depicted in
Figure 4.1.20 show the loss in capacity due to the guard band. For all the values of SNR, and all the
modulation orders, FBMC outperforms OFDM. Not only FBMC without guard carriers provides better
capacity than FBMC with guard carriers, FBMC with guard carriers gives a better capacity than OFDM
without guard carriers.
This difference in capacity is even further emphasized when the effect of a limited power control
feedback is measured on capacity. The effect of a power control mismatch has been simulated by
increasing the power of the asynchronous adjacent user. We depicted in Figure 4.1.21 the impact of
the power control on capacity.
FP7 ICT GA 318555
D3.3
Figure 4.1.21: Capacity in bit/s/carrier for OFDM and FBMC waveforms when power control feedback is limited
(assuming scenario (a)). Blue, red and green curves stand for respectively QPSK, 16-QAM and 64-QAM
modulations.
As previously mentioned, when QPSK modulation is considered the capacity is limited by thermal
noise level. Therefore even if the power of the adjacent user is increased by 3dB, the capacity
achieved with QPSK modulation is equal to the capacity in presence of perfect power control. The
impact of power control on capacity is more important for 16-QAM and 64-QAM modulation. For
FBMC, the impact of the adjacent user power is almost negligible. This is because only the carriers
located at the edges of the spectrum are affected by the interference. However for the case of
OFDM, the capacity is significantly reduced when the power of the adjacent user is increased. As the
interference level varies linearly with the power of the adjacent user, capacity decreases.
These results underline the benefit of the FBMC waveform in case of limited feedback for both time
misalignment and power control. We demonstrated the benefits of the FBMC waveform compared
to the OFDM waveform, particularly for high order modulations. Due to the fair frequency
localization, only the carriers located at the edges of the active spectrum are affected by
interference, while for OFDM interference is spread over all the active carriers. In case of fragmented
spectrum the impact of interference is further amplified for the case of OFDM. Moreover FBMC
FP7 ICT GA 318555
D3.3
waveforms permit a simple way of sharing resources between cell-edge users without strict
synchronization between users, especially in near unity frequency reuse. It is a direct consequence of
the low level of uplink interference generated by the built-in waveform filter.
4.2 Frame Structure and Multiple Access Performance
4.2.1 IDMA and Superimposed Pilots

In the previous Deliverable [D3.2] and literature [CSW14], we exhibit that a dual-user Interleave-
Division Multiple Access (IDMA) system outperforms traditional Frequency Division Multiple Access
(FDMA), if jointly combining with OFMD or UFMC:
The comparison was made in a setting with relaxed synchronicity requirements. According to Fig.
4.2.1, with low-rate forward error correction (FEC) codes, IDMA is strongly outperforming FDMA. In
order to compare the schemes under the same maximum throughput conditions, two 16-QAM FDMA
users occupy each half the allocation size as the two superimposed IDMA layers using QPSK. For
higher values of Eb/N0 the benefits of UFMC get more visible (e.g. comparing the green curves with
squares), especially in an asynchronous setting. The combination of UFMC with both multiple access
schemes is always better than pure CP-OFDM.
Here we see that the strengths of UFMC and IDMA are complementing each other well over different
ranges of SINR.
Fig. 4.2.1 Uplink OFDM vs UFMC using either FDMA or IDMA multiple access.
It is presented in [PLK06], that IDMA is capable of accommodating more than 40 users

simultaneously, even if a single antenna is available. Nevertheless, the prerequisite of good
performance is the Channel State Information (CSI) of multiple sub-channels available at the receiver
perfectly. This reveals the fact that channel estimation is extremely important for IDMA system, and
it is a big challenge to precisely estimate multiple sub-channels with a single element.
FP7 ICT GA 318555
D3.3
In this section, thus, we are motivated to focus on channel estimation issue for IDMA system within
an asynchronous environment. We exploit superimposed pilots as additional layers to the signal
layers, by allocating a marginal fraction of total transmission power. Basically, the superimposed
pilots are cyclically repeated user-specific spreading sequences, a priori known by corresponding
receiver. The superimposed pilots can provide an initial channel estimate by realizing the estimation
on the single-user basis with respect to the code-division orthogonality. Further, being similar to
[HT99] and [Ct13], the trellis-based channel estimator is embedded in the IDMA iterations, by dual-
purposely utilizing the superimposed pilots. Nevertheless, being unlike in [Ct13], we improve not
only the survival data sequence, but also the tracked channel iteratively. The so-called turbo-effect
can be obviously observed between two consecutive IDMA iterations, where the precision of data
detection and channel estimation are both enhanced.
4.2.1.1 Signal Model

In Fig. 4.2.1, the system diagram of an asynchronous IDMA system with superimposed pilots and
corresponding channel estimation is presented. Assume that N users are coexisting within the IDMA
system, by assigning a single layer to each user. In order to satisfy the low-rate characteristic of data
dn, signal of the n-th user, we concatenate a rate-Rc channel encoder with a rate-Rr repetition
encoder. The user data will be M-QAM modulated, after user-specifically interleaving dn. The
superimposed pilots pn in Fig. 4.2.1.1 is as well M-QAM modulated. The pilot of n-th user is basically a
cyclically repeated user-specific spreading sequence. The average data power per symbol and the
superimposed pilot power per symbol are denoted as Ed and Ep, respectively. It holds
Es Ed E p (4.2.1.1)
where Es denotes the energy per symbol.
In Fig. 4.2.2, we demonstrate how to achieve the code-division orthogonality by superimposed pilots
under asynchronous environment. Let us assume that each user adopts a pilot pattern as one
uniquely available Walsh Hadamard code p n , and cyclically repeats it. Hence, during the detection
window, an arbitrary pilot pattern is orthogonal to anyone of other pilot patterns, and the
corresponding orthogonality is independent to their timing advance or delay. Further, if the symbol
impulse function S(t) can be generated within interval [0, Ts], it holds,
4T s
0 p 2 k S ( t kT s ) p 4 k S ( t kT s lT s ) dt 0 (4.2.1.2)
k k
with k, l = 0, 1, 2, 3, and the variable ||<Ts indicates that the orthogonality can still be satisfied, if
the timing advance of delay is not an integer of multiple symbols.
Further, the system equation in Fig. 4.2.2 can be presented as

N
y hn E d rn E p
( pnw n ) z (4.2.1.3)

n 1
pn
where the vectors in (4.2.1.3) are all G1 dimensional, and G denotes the spread factor. The data
symbol vector rn is obtained by stacking the original data bits dn after encoding, repetition,
interleaving and modulation. The superimposed pilot vector is denoted by pn. Vector wn is a user-
specific spreading vector of the n-th user to spread a training symbol pn. Vector z represents the
Additive White Gaussian Noise (AWGN) with noise power spectral density N0.
FP7 ICT GA 318555
D3.3
Figure 4.2.2: Channel estimation for asynchronous IDMA with superimposed pilots
Figure 4.2.3: Satisfying the code-division orthogonality of superimposed pilots in an asynchronous environment
FP7 ICT GA 318555
D3.3
At the receiver, the despread-based linear channel estimator is deployed, which is able to remove
the superimposed pilot sequences of other users in the equation due to the provided orthogonality.
The initial channel estimates can be computed by

N
pnw
h n h n hi E d ri z .
(0) n
(4.2.1.4)
G E p i 1
The superscript of h n( 0 ) refers to the 0-th iteration as an initial channel estimation for IDMA detector.
The corresponding channel estimation Mean Square Error (MSE) can be represented by
NE N0
(4.2.1.5)
d
MSE .
G Ep
Notice that the MSE term in (4.2.1.5) is interference-limited, namely N0 does not play a dominant
role there. For example, the MSE is approximately -4.94dB, with N = 12 users, spread factor G = 64, Ed
= 0.6, Ep = 0.4, even at a Signal-to-Noise Ratio (SNR) value 1 0 log 10 ( E s N 0 ) 0 dB.
4.2.1.2 Detection and Estimation

In the previous deliverable [D3.2], the IDMA concept has already been well clarified. Thus, in this
section, let us focus on the trellis-based channel estimation. In [HL99] and [HT99], the per-survivor
processing and superimposed pilots based channel estimation are proposed, respectively. In [Ct13],
we realize the trellis-based joint data detection and channel tracking for MIMO communications.
Nevertheless, channel tracking is different from channel estimation, because the tracked channel is
not exploited by the detector or decoder to enhance the iterative processing. Hence, in this section,
we exploit superimposed pilot to jointly track/predict the channel and detect the signal, by
considering per-survivor processing as well. Being unsimilar to previous works mentioned, we store
the tracked channel, exchange the channel estimation and the decoded data between the IDMA
detector/decoder and trellis-based channel estimator. This allows us having a very precise channel
estimation, even at very low SNR.
Without the loss of generality, we consider a trellis with memory length = 2 with BPSK modulation,
throughout this section. As depicted in Fig. 4.2.4, at an arbitrary time instant k, anyone of the four
possible states comprises two previously transmitted bits bk-1 and bk. The state tuple (bk-1, bk)
corresponds to a state ID Sk. For the convenience, the user index n is omitted here. At time instant k,
the corresponding system equation with the memory is
y k i hk i ( E d s k i E p p k i ) n k i . (4.2.1.6)
with i = 1, , = 2. The channel can be straightforwardly reconstructed as
E d s k i E p p k i
* *
~
hk i y k i hk i n k i .
r
2
(4.2.1.7)
E d s k i E p p k i
with E [ n~ k i ]
2
E s N 0 (2 E s )
2
. Thus, the channel for time instant k can be predicted by

1
hk w k i hk i
p r r
hk i (4.2.1.8)
i 1 i 1
where the prediction coefficients are denoted as wk-i. Although the optimal prediction corresponds to
Wiener extrapolation, we simply average the reconstructed channels by adopting wk-i = 1/, because
it achieves identical performance as Wiener approach with typically small memory . Let us compute
the Euclidean distance of the transition, namely the branch metric, as
FP7 ICT GA 318555
D3.3
2
r
1 hk i
k
yk ( E d sk E p
p k ) . (4.2.1.9)
N 0 i 1
Figure 4.2.4: Trellis structure for BPSK modulated signal
With the trellis structure, the Bahl-Cocke-Jelinek-Raviv (BCJR) algorithm [BCJ74] can be deployed to
recursively detect the signal in a Maximum a posteriori (MAP) manner, by introducing the forward
recursion coefficients , the backward recursion coefficient , and the branch metric k described
above jointly. We notice that the most likely channel prediction will be stored for each Current-State,
referred to as local survivor. This makes it possible to pick out the most likely channel estimation
sequentially, after the forward and backward recursions within BCJR algorithm, referred to as global
survivor. For IDMA detector/decoder, we will exploit this global survivor as the channel estimates
and improve itself throughout the IDMA iterations.
4.2.1.3 Iterative Processing

In this paragraph, we will focus on the technique details of combing IDMA detector/decoder with
trellis-based channel estimator. A short summary with respect to the iterative processing will be as
well provided. Being a part of Fig. 4.2.2, Fig. 4.2.5 involves uniquely the iterative processing for user
n.
FP7 ICT GA 318555
D3.3
Figure 4.2.5: Iteratively estimate the channel and decode the data for n-th user
In previous section, equation (4.2.1.4) demonstrates how to produce the initial channel estimate
h n for user n with superimposed pilots, even suffering from possible asynchronicity. It enables to
(0)
( out )
generate the soft-out L ESE ,n
with respect to the operation in Elementary Signal Estimator (ESE).
( in )
From the first iteration on, the trellis detector starts to play a role. The re-interleaved soft bits L ESE ,n
serve as the a priori information for the trellis detector. After cancelling out the Multiple Access
Interference (MAI), consisting of remaining data signal and superimposed pilots, the observation yn is
given by
h l
( m 1 )
yn y ( E d rl E p
pl ) . (4.2.1.10)
ln
The tracked channel h n( m ) of m-th iteration will then serve as the channel estimate for ESE. On the
( out )
other side, the trellis detector delivers in parallel the soft-out L TD , n for the data sequence. The
( out ) ( out )
likelihood can be improved by combining L TD , n and L ESE ,n
. As the iterative processing cannot be
avoided in IDMA system, the trellis detector is the only factor to increase the complexity.
Nevertheless, by selecting low order modulation scheme, e.g. BPSK, the additional computational
requirement is absolutely affordable.
4.2.1.4 Simulation and Numerical Results

In this section, we focus on point-to-point link level simulation to evaluate the performance. In order
to prove the concept, we adopt an IDMA system with N = 4 users and BPSK modulation. At the initial
stage, in order to achieve acceptable initial channel estimate h n( 0 ) , we exploit a relatively bigger
spreading factor G, and the normalized channel fading rate Ts fD is assumed to be 10-3. It corresponds
to a pedestrian with approximately 3km/h velocity, with respect to the LTE system at 2GHz frequency
band with symbol rate Ts = 66.67s. In Table 4.2.1, the parameters of link level simulation are
summarized.
In Fig. 4.2.6 and Fig. 4.2.7, a snapshot for the channel estimation is provided with respect to its
magnitude and phase. We compare basically the initial channel estimation h n ( t ) to the channel
(0)
estimation after 2nd IDMA iteration h n ( t ) . The SNR Es/N0 is -2dB with further parameters Ed = 0.9, Ep
(2)
= 0.1.
FP7 ICT GA 318555
D3.3
Table 4.2.1: Link level simulation parameters
The effective Eb/N0 for M-QAM can be computed as

Eb Ed 1 1
. (4.2.1.11)
N0 N0 log 2
M Rc Rr
Figure 4.2.6: Magnitude: compare h n ( t ) with h n ( t ) for n-th user, Ed = 0.9, Ep = 0.1, Es/N0 = -2dB, Ts fD = 10
(0) (2) -
3
, 4 users, BPSK
According to the example, the Eb/N0 is about 9.6dB. If the perfect channel information is available,
the IDMA receiver is able to detect/decode the user data successfully. On the other side, the trellis
detector works on the Es/N0 basis, which is extremely low in this case. For this reason, we are going
to check, whether the channel can be precisely estimated. In both Fig. 4.2.6 and Fig. 4.2.7, notice that
the initial channel estimation can roughly, but not precisely track the channel. The LDPC channel
decoder can anyway improve the bit error and deliver the soft-out values to the trellis detector as a
priori information. It is illustrated in both Fig. 4.2.6 and Fig. 4.2.7, that the multiple channels of 4
users can be precisely tracked only after two iterations, at very low SNR Es/N0 = -2dB.
FP7 ICT GA 318555
D3.3
Figure 4.2.7: Phase: compare h n ( t ) with h n ( t ) for n-th user, Ed = 0.9, Ep = 0.1, Es/N0 = -2dB, Ts fD = 10 , 4
(0) (2) -3
users, BPSK
Figure 4.2.8: Channel estimation MSE vs. Es/N0: compare h n with h n for n-th user, Ts fD = 10 , 4 users,
(0) (2) -3
BPSK
In Fig. 4.2.8, the channel estimation MSE versus SNR Es/N0 is presented, with respect to h n( 0 ) , h n
(2)
and diverse setting for Ed, Ep. Surprisingly, we notice that the parameters setting Ed = 0.9, Ep = 0.1
turns out to be the most effective approach to optimize the MSE, even if the quality of initial channel
estimation h n( 0 ) is still poor. On the other side, this indicates that trellis-based channel estimator is
very robust and powerful to track the channel, even if the energy of superimposed pilots is marginal.
FP7 ICT GA 318555
D3.3
Figure 4.2.9: Channel estimation MSE vs. Es/N0: compare h n , h n with h n

(0) (1 ) (2) -3
for n-th user, Ts fD = 10 , 4
users, BPSK
In Fig. 4.2.9, the similar results are presented for channel estimation MSE versus SNR Es/N0. Fig. 4.2.9
tries to demonstrate the convergence of the MSE of channel estimation, after the first and second
iteration. It can be clearly observed, that the rate of convergence is very satisfactory, even if the
energy of the superimposed pilots is low, e.g. with Ep = 0.1.
4.2.1.5 Summary
Potentially being one important component of future wireless standard, the superimposed pilots
based techniques are convincing the people, and attracting more and more attentions, for its
capability of delivering robust performance without having to sacrificing users information rate. In
this section, we firstly exhibit that the superimposed pilots can be integrated to an IDMA system to
provide an initial channel estimation, even with a marginal energy per superimposed pilot symbol.
Further, the initial channel estimation can be obtained under the asynchronous environment,
because the code-division orthogonality can be protected by the designed superimposed pilots.
Secondly, we demonstrate that the roughly obtained initial channel estimation can be further
improved, by introducing a trellis detector. By iteratively exchanging the soft-out values between the
channel decoder and trellis detector, the multiple IDMA sub-channels can be precisely tracked, even
with only two iterations.
FP7 ICT GA 318555
D3.3
4.2.2 D-PRACH Performance
4.2.2.1 Theoretical analysis

In the course of the project we have derived a thorough interference analysis [KWJ+14, WKJ15,
WGA15, WKJ+15]. The main results are as follows:
Our system model has to capture that many users each occupy a small number of subcarriers and
each of them asynchronously (in time and frequency or both) access this resource in an
uncoordinated fashion. Apparently, there is no proper analytical model to handle such asynchronous
random access.
For a particular time-frequency slot we will denote the (random) channel operator
as and the asynchronism as . We assume that can be estimated using channel
estimation procedure while cannot (only by its mean). Writing the received complex symbol
in the absence of additive noise yields
, (4.2.2.1)
where we defined , i.e. the mean value conditioned on a fixed channel 1. Thus
the transmitted symbol will be multiplied by a constant and disturbed by two zero mean random
variables (RV), and ICI. The first RV represents a distortion which comes from the
randomness of , the second term ICI represents both. The mean power of both contributions,
conditioned on a fixed channel , are where
and Each element of the distortion sum is given by:
(4.2.2.2)
The objective here is to analyse the distortion terms and over all subcarriers.
Theorem 1 Suppose (without loss of generality) such that
(perfect reconstruction in noiseless case). The average distortion power per subcarrier is
upperbounded by
(4.2.2.3)
where:
.
As a special case, assume now a deterministic time-frequency shift . This distortion
is non-random and energy preserving, i.e. . Evaluating the function gives:
(4.2.2.4)
Hence, in the case of AWGN case we have:
1
As a matter of fact, the expectations depends only on the marginal distribution of .
FP7 ICT GA 318555
D3.3
(4.2.2.5)
Let us now investigate the influence of time and frequency offsets on the spline-based waveform.
In order to obtain a bound on the influence of time or frequency offsets we need to evaluate the
cross-ambiguity function . For this the dual pulse has to be taken into account. However, since
this is not desirable, simple estimates of this function are needed. The following theorem provides a
bound on [WKJ15].
Theorem 2 Suppose (without loss of generality) such that
(perfect reconstruction in noiseless case). then
(4.2.2.6)
where denotes the Fourier transform of .
4.2.2.2 Numerical evaluation

Subsequently, we show using numerical experiments that the 5GNOW D-PRACH concepts in
combination with BFDM is feasible. We compare the standard (LTE) PRACH implementation to our
proposed spline based pulse-shaped PRACH. The main simulation parameters, chosen according to
LTE specifications, are provided in Table 4.2.2. Let us spend a few more words on the simulation
concept.
PUSCH Standard Pulse shaped

PRACH PRACH
Bandwidth 20 MHz 1.08 MHz 1.08 MHz
OFDM symbol 0.67 s 800 s -
Subcarrier spacing F 15 kHz 1.25 kHz 1.25 kHz
Sampling frequency fs 30.72 MHz 30.72 MHz 30.72 MHz
Length of FFT NFFT 2048 24576 24576
Number of subcarrier L 1200 839 839
Cyclic prefix length Tcp 160Ts 1st 3168Ts 0
144Ts else
Guard time Tg 0 2976Ts 0
Pulse length P - - 4 ms
Number of symbols K 14 1 1
Time-freq. product TF 1.073 1.25 1.25
TABLE 4.2.2: System Specification
For the computations involved, we make use the LTFAT toolbox, which provides an efficient
implementation. Due to the properties of the pulses and to fit the strict LTE frequency specification,
we allow a small spillover effect from PRACH to PUSCH in time. Moreover, due to the PRACH pulse
length of 4 ms, we simulate the PUSCH over this time interval. Furthermore we use the maximal
available LTE bandwidth of 20 MHz. In the LTE standard, the power of PRACH is variable and is
incrementally increased according to a complicated procedure. To allow a meaningful comparison
(without having to implement the complete PRACH power ramp-up procedure) we choose the power
of the PRACH such that approximately the same power spectral density as in PUSCH is achieved. This
is depicted in Figure 4.2.10.
FP7 ICT GA 318555
D3.3
Figure 4.2.10: Power spectral density. The power of the PRACH is chosen to achieve a PSD
similar to PUSCH
We simulate ideal channels in which the only distortion comes from frequency offsets and time
delays, which are the main challenges for the system. Moreover, we assume a maximum length of
the channel impulse response of 300 samples, which corresponds to a delay spread of roughly 5 s,
and which implies a maximum cell radius of 1.5 km. The following results are obtained using BPSK
modulation.
4.2.2.3 Influence of asynchronicity on single-user PRACH

Subsequently, we consider only simulation results for the data part of the PRACH, hence we do not
consider the PUSCH in order to keep the simulation as simple and meaningful as possible. In the
same spirit, the control PRACH also remains inactive in order to compare the interference between
data-subcarriers without being subject to any external effect. More advanced simulations are
presented in [WKJ+15].
In Figure 4.2.11, we show the effect of time delays on the normalized interference considering a data
transmission on a variable number of D-PRACH subcarriers (20 and 2000) for a single user. Both the
standard LTE PRACH and the new spline-based PRACH approach are compared. The results reveal
that the interference of the spline-based approach remains significantly lower than the OFDM-based
method for negative offsets, while it gets slightly higher for small positive shifts inside the cyclic
prefix. In addition to this, we observe that the theoretical interference bounds are met quite well
hence proving the applicability of our theoretical analysis.
FP7 ICT GA 318555
D3.3
Figure 4.2.11: Simulated interference and interference bound vs. negative and positive time
offsets for OFDM and spline waveform for different numbers of subcarriers.
Figure 4.2.12: Comparison of OFDM and spline waveforms in terms of interference as well as their
respective interference bounds. Numerical results of different numbers of subcarriers are shown.
FP7 ICT GA 318555
D3.3
In Figure 4.2.12, we show the effect of frequency offsets on the normalized interference for a data
transmission on different numbers of D-PRACH subcarriers for a single user. Again, the standard LTE
PRACH OFDM-based method and the proposed spline-based PRACH approach are compared. We
observe, that the interference remains significantly lower for the spline-based approach (comparing
equal numbers of data-subcarriers). Moreover, the theoretical interference bounds are, as in the
previous case, never exceeded.
Clearly, the performance of PRACH does not deteriorate due to the proposed spline-based PRACH. By
contrast, irrespective of the actual number of subcarriers used for data transmission, the spline-
based approach leads to an overall slightly reduced interference level for the PRACH data-
subcarriers.
4.2.2.4 Asynchronous multiple users

Asynchronous data transmission is a major challenge that comes with MTC and the Internet of
Things. Therefore, we now consider a second, completely asynchronous, user that transmits data in
the PRACH. Thereby we assign half of the subcarriers available for PRACH data transmission to this
second user by subdividing the PRACH into continuous blocks of neighboring subcarriers that are
symmetrical around the control-PRACH. However, we still evaluate only the performance of the
original user of interest (and consequently we decode only this user), which is assumed to transmit
at the inner subcarriers close to the control PRACH. Thereby, we compare two waveforms, OFDM
and the proposed spline approach. This user has a small fixed time delay (4 s) below the cyclic prefix
(103 s), whilst the second user has a variable delay that can exceed the cyclic prefix duration.
Figure 4.2.13: Simulated interference for OFDM and spline waveforms in presence of an
asynchronous interfering user with varying time delay.
The results are depicted in Figure 4.2.13. It can be observed that for time delays larger than the cyclic
prefix length, the interference of the OFDM-based system grows vastly while the interference of the
spline-based system increases in a slower fashion. Furthermore, the interference for negative delay
values also increases far slower in the spline-based modulation in comparison to the OFDM-based
FP7 ICT GA 318555
D3.3
system (for which the interference is significantly higher in this range). In order to illustrate the
connection to the previous results and the theoretical bounds, we show that the simulated
interference for the two-user case approaches the interference for a single user, when the difference
in the offsets becomes smaller.
Figure 4.2.14: Simulated interference vs. Time offset for the spline waveform. Numerical results for the case of
20 subcarriers are shown. In addition, the behavior of the two-user case is illustrated, where each user has a
different delay. The (aggregate) interference converges to the interference value for a single user at a certain
delay value when the delays of both users tend towards this very same delay value.
Figures 4.2.14 and 4.2.15 show that for a two-user scenario with relatively close time delay- or
frequency offset values, the resulting interference remains close to that of a single-user experiencing
the mean of these values. Hence, it is interesting to observe that the interference value of a single-
user at the mean of two time delays or frequency offsets affected to two users can be considered as
the minimal interference that can be obtained in the two-user case.
Figure 4.2.16 extends the results shown in Figure 4.2.13 to symbol error rates. Similar to the
interference, the SERs remain significantly lower for negative time delays (as well as for delays higher
than the cyclic prefix) in the spline case.
Figure 4.2.17 shows the same performance where one user has an additional frequency offset. The
performance of BFDM is now significantly better than OFDM particularly when we consider smaller
symbols and symmetric distortion where the receive window is symmetric with respect to zero delay.
FP7 ICT GA 318555
D3.3
Figure 4.2.15: Simulated interference vs. Frequency offset compared to the theoretical bound for the OFDM
waveform. Numerical results for the case of 20 subcarriers are shown. In addition, the behavior of the two-
user case is illustrated, where each user has a different frequency offset for a number of 2000 subcarriers. The
(aggregate) interference converges towards the value of the 2000 interference single-user curve (which
happens to lie very close to the bound) when the difference between the two offsets, which are centered
around 0.2 1250 Hz, becomes smaller.
Figure 4.2.16: Simulated symbol error rate vs. varying time delay of an asynchronous interfering user for the
OFDM and spline. The simulation scenario includes a channel estimation and both PRACH and PUSCH-signals
are sent. Numerical results for the case of 20 subcarriers are shown.
FP7 ICT GA 318555
D3.3
Figure 4.2.17: Symbol error rate in PRACH (using 4QAM) with perfect channel knowledge, averaged over 10 out
of 20 data subcarriers vs. a varying time offset of a second user with a frequency offset of 62,5 Hz. The SNR is
set to 25 dB. The black line shows the CP length in LTE PRACH.
FP7 ICT GA 318555
D3.3
5 Comparison and Discussion of Selected Waveforms
The 5GNOW requirements defined in [D2.1] target for reduced control overhead and increased
throughput compared to an LTE(-A) system. The 5GNOW waveforms selected by WP3 enable an
operation with relaxed synchronicity (see section 3.3 for results for UFMC and FBMC). This relaxed
synchronicity allows shortening the LTE access procedure and/or using contention-based access
without requiring closed-loop timing control. The overall system level gains in terms of overhead
reduction are addressed in [D4.2].
Compared to OFDM, all waveform candidate technologies clearly improve the out-of-band radiation
and fulfill the 5GNOW WP2 target of a factor 100 improvement in this KPI. The 5GNOW scenario
fragmented spectrum will thus be better supported. The target of supporting real-time constraints,
thus lower latency, can be solved by respective waveform numerologies, as given in the candidate
parameter sets in section 4.1. The evaluation of the waveforms in the context of the 5GNOW
scenarios with respect to 5GNOW KPIs is done in [D2.2, D2.3].
Altogether 5GNOW proposes a multicarrier Gabor-like structure for 5G coming in different variants
to be operated possibly in parallel in the unified frame structure and optimized parameter settings.
We have the following findings:
1) GFDM allows a flexible parametrization. In its default parameter set it is non-orthogonal and
aims at very short frame durations for low latency support. 5GNOW has extended the
theoretical basis for such systems allowing for transmission at critical density even if the
(continuous-time) Balian-Low theorem is violated. The reason is the block cyclic structure
(similar to the BFDM concept) which makes it possible to exploit discrete artefacts in the
concept. Notably, the non-orthogonality requires possibly advanced receivers for removing
self interference.
2) UFMC is very close to OFDM in its design with (quasi-)orthogonal reception, which allows to
directly apply all OFDM know-how from MIMO, channel estimation etc. Its key difference to
the other waveform candidates is, instead of applying a per-subcarrier filtering, entire groups
of subcarriers are filtered. The motivation is that ICI occurs between groups of subcarriers.
This allows to shorten the filter for efficiency and latency reasons.
3) FBMC in its default parameter setting has the strongest spectral localization of the waveform
candidates. This comes along with long filter durations and usage of offset-QAM for
efficiency reasons. FBMC makes use of a very advanced theory of Wilson bases allowing
orthogonal transmission even for systems operating at the critical density.
4) BFDM has a pulse design which is optimized for sparsity-aware detection techniques, making
it appealing for advanced receivers for D-PRACH. BFDM allows for quasi-OFDM like
operation with bi-orthogonal pulses (thereby giving up transmit orthogonality) so that many
OFDM concepts can be transferred but sensitivity to temporal and spectral asynchronisms is
dramatically improved particularly in the context of ATA.
Table 5.1 provides a comparative overview on waveform properties, including references to the
different assessed properties. As expected, there cannot be made a simple statement which is the
FP7 ICT GA 318555
D3.3
better waveform as this depends on the overall system context where the waveform is embedded
in and the respective system targets. This holistic 5G system design is a task which will be addressed
in future projects, e.g. 5GPPP. Those projects can built upon the 5GNOW frame structure and
waveform approaches in conjunction with optimized/reasonable parameters provided in chapter 4.
FP7 ICT GA 318555
IR2.1
UFMC FBMC GFDM BFDM

Orthogonality Complex domain Real domain Orthogonal / non- Bi-orthogonal
orthogonal / real
domain orthogonal
(sec. 4.1.1.3)
QAM/OQAM QAM OQAM QAM and OQAM QAM
preferred?
BER in multipath Slightly better than OFDM (no Slightly better than OFDM (no - Potential to Improved compared
Rayleigh energy wasted for CP). energy wasted for CP). reduce amount of to OFDM due to better
Additional advantages in Additional advantages in CP in the signal adaption to channel
scenarios with ICI. scenarios with ICI. (energy savings in profile [JW07]
the order of 7-
15% )
Non-orthogonal
parameterization
can cause self-
interference and
hence increase BER
compared to
orthogonal
schemes in order to
provide benefits
with other KPIs.
[MMI+14]
PAPR Similar as OFDM. Similar as OFDM. Can be configured Similar to OFDM: even
Allows also for single carrier for PAPR reduction complex methods can
variant: UFMC-based SC-FDMA beyond SC-FDMA be transferred [GW15]
possible with e.g. DFT- and easy
precoding as in OFDM (any integration with
technique can be carried over) any PAPR reduction
technique available
for OFDM. [MF13]
Out of band Arbitrary low side lobe level Arbitrary low side lobe level Can be configured Better than OFDM,
radiation floor due to subband filters thanks to per sub-carrier for a desired OOB comparable to GFDM
(spectrum fig. 2.3.2) filtering. Up to -60dB suppression at the but worse than FBMC
attenuation at adjacent cost of other KPIs. due to block cyclic
subcarriers (see fig. 2.4.2). [MMI+14] opertation [KWJ14]
Throughput Better than OFDM (smaller CP removal increases capacity In the order of 25% Better than OFDM
guards). compared to OFDM but filter better than OFDM: particularly for higher
raise and fall time must be - Smaller guard Doppler speeds and
taken into account. bands channel profile
Throughput is then slightly Less CP adaption [JW07]
better than OFDM for long
frames.
Spectral Efficiency Better than OFDM (smaller Can be better than OFDM, (see above) Better than OFDM
guards). 10% more usable depending on the overlapping particularly for higher
resource elements (fig. 2.3.5) factor, the number of OFDM Doppler speeds and
symbols and the size of the CP channel profile
(see Eq. (4.1.3.6)). adaption [JW07]
Complexity: Efficient frequency domain Efficient frequency domain Efficient frequency Logarithmic
Modulator / FFT-based implementation. IFFT/FFT-based domain DFT-based complexity growth in
Demodulator Slightly higher than OFDM due implementation, but still implementation block length is
to filtering higher complexity than OFDM available. preserved
due to filtering (see [D5.2]). Complexity within
in the order of 2-
10x times OFDM
complexity due to
filtering and
optional self-
interference
cancellation.
[GAS13], [MGK+12]
Complexity: As low as OFDM On the same order of Same as OFDM in Logarithmic
Remaining inner magnitude compared to case of no iterative complexity growth in
receiver until FEC OFDM (see [D5.2]). interference block length is
input cancellation preserved
Training structure OFDM know-how can be OFDM know-how can be re- OFDM know-how Quasi-OFDM like
and properties reused. All degrees of freedom used but need adaptation for can be reused. operation possible
available for complex channel estimation and OQAM and/or non- [KWJ+14]
FP7 ICT GA 318555
IR2.1
sequences with multi-user MIMO. Burst modes with orthogonality can

multiplexing. preambles allows time- be compensated.
frequency synchronisation [GMM+14]
and channel estimation.
Channel estimation OFDM know-how can be OFDM know-how can be OFDM know-how Quasi-OFDM like
reused. Low complex reused. Low complex can be reused. operation possible
frequency domain channel frequency domain channel OQAM and/or non- [KWJ+14]
estimation fully supported. estimation fully supported orthogonality can
(see section 4.1.3) when be compensated.
interference cancellation [MMF15]
methods are applied at the
transmitter side (a more
complex frame structure is
required )
Sensitivity to time- Performance gains over OFDM Performance gains over OFDM Performance gains Major advantage of
freq. offsets shown by fig. 3.3.1 (see figure 4.1.3.9 and section over OFDM can BFDM compared to
4.1.3) achieve 20dB OFDM: order of
[MMF15]. magnitude better
[KWJ+14,WKJ+15]
Multiuser capability Block-based modulation Block-based modulation Easy integration Major advantage of
configurable in time and configurable in time and with LTE BFDM compared to
frequency, handling mix of frequency. Support for multi- approaches. OFDM particularly in
sync./async. users as discussed user (MIMO) schemes with [GMF+14] asynchronous uplink
in section 3.3 adaptation, handling mix of Full support MU- [KWJ+14,WKJ+15]
Full support for multiuser sync./async. Users (see figure MIMO-GFDM.
(MIMO) multiplexing of 4.1.3.10). [MLG+15]
sounding/small control
elements
Synergy with IDMA shown in
section 3.2
Multipoint capability Performance gains over OFDM Performance gains over Open topic No specific advantage
shown by fig. 2.3.6 OFDM: avoid multiple access of BFDM in
interference in uplink (see synchronous
figure 4.1.3.10 and 4.1.3.11), operation
less complex than equivalent
CP-OFDM with MAI-
cancellation, better than
OFDM for given scenario with
impairments (e.g. precision of
local oscillators, see section
4.1.3).
Resource allocation As with OFDM Same as OFDM Two-dimensional Quasi-OFDM like
mechanisms / data block operation possible
channel adaptive structure enables [KWJ+14]
scheduling more degrees of
freedom regarding
resource allocation
than OFDM.
[MMI+14]
Adaptivity potential Subband-specific waveform Sub-carrier/Sub-band specific Extremely flexible Better than OFDM
parameters possible waveform parameters waveform with a particularly for higher
possible. Flexible parameters: large number of Doppler speeds and
number of symbols per block, degrees of freedom channel profile
number of subcarriers, allows on the one adaption [JW07]
hand to consider
classical OFDM and
SC-FDE as corner
cases and on the
other hand to
explore parameter
sets optimized for
specific scenarios,
e.g. low latency
communication.
[MMI+14]
Coexistence to Good due to low out-of-band Very good due to very low - Can be Similar to GFDM
legacy systems radiation. Can re-use LTE out-of band radiation. Can re- parameterized for
parameter settings, signaling use LTE parameter settings optimal
and pilot symbols and part of the signaling and protection of
FP7 ICT GA 318555
IR2.1
pilot symbols. legacy frequency

bands in
asynchronous
operation mode
[MLR+11]
Can be
parameterized to
reuse the LTE
master clock and
resource grid.
[GMF+14]
Key open problems UFMC based air interface FBMC based air interface - Complexity for Tradeoff between
design not yet fully completed design not yet fully large K and M transmit/receive
completed. Need further - Synchronization orthogonality and
investigation for (MU) MIMO. under severe time/frequency
frequency offsets localization
- Improve
performance with
low-cost front
ends
- Spatial
multiplexing of
large number of
users
Optimal receiver
for spatial
multiplexing
(sphere detection)
Test environments Relaxed synchronicity, short Relaxed synchronicity, DL - Low latency one Not available
burst communication, UL CoMP, fragmented spectrum shot transmission
CoMP, fragemented spectrum Also: MTC, bitpipe
and spectrum
aggregation,
WRAN, V2V
communication
[MMI+14]
Intrinsic motivation User allocations are happening Fair frequency localization of Opening new Asynchronous
for in groups of subcarriers. This the prototype filter perspectives for operation
can be done for filtering as well guarantees that only adjacent using 5G networks
to reduce filter length. carriers interfere with each in unprecedented
Keep all OFDM advantages, but others: great motivation for application
improve spectral properties non-synchronous context scenarios.
and benefit from this. (incl. CoMP) and fragmented
Consensus in 5G spectrum.
standardization possibly easy
to achieve.
FP7 ICT GA 318555
IR2.1
6 Conclusion
LTE and its evolution LTE-A are standardized via the 3rd generation partnership project (3GPP) [23].
The foreseen diversification of the service and device-class mix of future telecommunications and the
related expansion of the requirement space [WJK+14] require a revolutionary step. This step from 4G
to 5G, anticipated in the 5GNOW project goals, implies a backward compatibility drop. 5GNOW new
waveforms and the usage of the 5GNOW unified frame structure with a mixture of synchronous and
asynchronous traffic are a major building block for supporting those goals. Starting from the main
drivers of 5G, sporadic traffic, spectral and temporal fragmentation and real-time constraints
together with the vision of supporting a single unified air interface, D3.3 presents such building
blocks. General waveform considerations are discussed in the light of the Gabor theory. Then the
four 5GNOW waveform candidate technologies GFDM, UFMC, FBMC and BFDM are described in
detail, summarizing the available results. These waveforms are supporting and enabling the unified
frame sturcture. The unified frame structure concept is the heart of the 5GNOW frame design,
designed for supporting various heterogeneous traffic and device types in parallel. We argue that
each multicarrier waveform has properties which makes it advantageous in specific scenarios and
shall be operated in parallel with others by proper re-parameterization. Performance results for the
candidate transceiver approaches are provided, including optimized and/or reasonable waveform
parameters.
Those new signal formats require standardization, as they need to be known on both ends of the link.
The 5GNOW project assesses the advantages gained when using the new 5GNOW technologies and
thus will generate technical findings which guide the decisions on this generation change. 5GNOW
has provided a powerful waveform and frame structure toolbox for the standardization process,
coming along with optimized/reasonable parameter settings and performance results. The further
system design steps taken in future 5GPP projects can built upon this vast number of available
5GNOW technologies and results for waveforms and frame structure of a new 5G air interface,
paving the way for 5G standardization.
The wireless industry as a whole has to build up consensus on the technology candidates for 5G
standardization. For this purpose, 5GNOW is in close contact to the European METIS research project
[24], in order to spread 5GNOW outcomes on a broad basis into the industry and research
community. The encouraging results, 5GNOW and METIS have achieved so far, lay the ground for the
arising 5G infrastructure PPP projects [25]. Those projects, based on generated 5GNOW know-how,
guided by METIS system concepts, will then be able to directly work towards pre-standardization.
3GPP release 14, starting in 2016 could be a first platform for creating a study item focused on a new
air interface.
FP7 ICT GA 318555
IR2.1
7 Abbreviations and References
3GPP 3rd Generation Partnership Project

4G Fourth Generation
5G Fifth Generation
5GNOW 5th Generation Non-orthogonal Waveforms for Asynchronous
Signaling
ADC Analog-to-Digital Converter
BER Bit Error Rate
BS Base Station
CoMP Coordinated Multipoint
DAC Digital-to-Analog Converter
EXALTED Expanding LTE for Devices
FBMC Filter Bank Multi-Carrier
FP7 7th Framework Programme
GFDM Generalized Frequency Division Multiplexing
GSM Global System for Mobile Communications
H2H Human-to-Human
ICI Inter-Carrier Interference
KPI Key Performance Indicato
LTE Long Term Evolution
LTE-A Long Term Evolution Advanced
MAC Medium Access (layer)
MTC Machine Type Communication
FP7 ICT GA 318555
IR2.1
References:
[3GP09] http://www.3gpp.org/ftp/workshop/2009-12-17_ITU-R_IMT-Adv_eval/docs/pdf/REV-
090006.pdf, last access 13.11.2012
[3GP10] 3GPP TR 36.814, Further Advancements for E-UTRA Physical Layer Aspects, v9.0.0,
Mar. 2010
[3GP11] 3GPP TR 36.819 V11.1.0, Coordinated Multi-point Operation for LTE Physical Layer
Aspects, December 2011
[3GP11a] Technical Specification 22.368 V11.2.0 (2011-06) Stage 1 (Release 11),
http://www.3gpp.org/ftp/Specs/html-info/22368.htm
[5GPP] 5GNOW Project Proposal, Part B
[BCJ74] Bahl, L.R.; Cocke, J.; Jelinek, F.; Raviv, J.; Optimal decoding of linear codes for
minimizing symbol error rate, IEEE Trans. Inf. Theory, Vol. 20, No. 2, pp. 284287,
Mar. 1974.
[BDN14] V. Berg, J-B. Dor, and D. Noguet, A flexible FS-FBMC receiver for dynamic access in
the TVWS, 9th International Conference on Cognitive Radio Oriented Wireless
Networks and Communications (CROWNCOM).
[Bel10] M. Bellanger et al., FBMC physical layer: a primer, June 2010. [Online] Available:
http://www.ict-phydyas.org
[Bel12] M. Bellanger, FS-FBMC: A flexible robust scheme for efficient multicarrier
broadband wireless access, in 2012 IEEE Globecom Workshops, Dec. 2012, pp. 192-
196.
[Bin90] J. Bingham, Multicarrier Modulation for Data Transmission: An Idea Whose Time
Has Come, IEEE Communications Magazine, vol. 28, no. 5, pp. 514, May 1990.
(VTC Fall12), Qubec City, Canada, Sep. 2012, pp. 15.
[Bol01] H. Bolcskei, Blind estimation of symbol timing and carrier frequency offset in
wireless OFDM systems, IEEE Transactions on Communications, vol. 49, no. 6, pp.
988999, 2001.
[BTS+13] Berardinelli, G.; Tavares, F.M.L.; Sorensen, T.B.; Mogensen, P.; Pajukoski, K., "Zero-
tail DFT-spread-OFDM signals," Globecom Workshops (GC Wkshps), 2013 IEEE,
pp.229,234, 9-13 Dec. 2013.
[Chu72] D. Chu, Polyphase codes with good periodic correlation properties (corresp.), IEEE
Transactions on Information Theory, vol. 18, no. 4, pp. 531532, 1972.
[Cis10] Cisco, "Architectural Considerations for Backhaul of 2G/3G and Long Term Evolution
Networks", White Paper, 2010 [Available: http://www.cisco.com/en/US/solutions
/collateral/ns341/ns973/white_paper_c11-613002.pdf]
[CKD13] N. Cassiau, D. Ktnas, and J-B. Dor, "Time and frequency synchronization for CoMP
with FBMC", Wireless Communication Systems (ISWCS 2013), Proceedings of the
Tenth International Symposium on, 27-30 Aug. 2013, Ilmenau, Germany.
[CKK12] O. Christensen, H. O. Kim, and R. Y. Kim, Gabor windows supported on [-1,1] and
dual windows with small support, Advances in Computational Mathematics, 2012.
[CSW14] Chen, Y.; Schaich, F.; Wild, T.; Multiple Access and Waveforms for 5G: IDMA and
Universal Filtered Multi-Carrier, in proceedings of IEEE Veh. Technol. Conf. Spring
(VTC14 Spring), Seoul, Korea, May 2014.
[Ct13] Chen, Yejian; ten Brink, S.; Pilot Strategies for Trellis-Based MIMO Channel Tracking
and Data Detection, in Proc. of IEEE Global Telecommun. Conf. (Globecom13), pp.
43134318, Dec. 2013.
FP7 ICT GA 318555
IR2.1
[D2.1] 5GNOW deliverable D2.1, 5G Cellular Communications Scenarios and System

Requirements, 22/12/2014
[D2.2] 5GNOW deliverable D2.2, Consistent 5G Radio Access Architecture Concepts,
21/11/2013
[D3.1] 5GNOW deliverable D3.1, 5G Waveform Candidate Selection, 21/11/2013
[D3.2] 5GNOW deliverable D3.2, Intermediate 5GNOW Transceiver and frame structure
concept, 14/03/2014
[Dau92] I. Daubechies, Ten Lecures on Wavelets, Philadelphia, PA: SIAM, 1992.
[Dem98] T. Demeechai, Pulse-shaping filters with ISI-free matched and unmatched filter
properties, IEEE Transactions on Communications, vol. 46, no. 8, p. 992, 1998.
[Dha01] N. Al-Dhahir, Single-Carrier Frequency-Domain Equalization for Space-Time Block-
Coded Transmissions over Frequency-Selective Fading Channels, IEEE
Communications Letters, vol. 5, no. 7, pp. 304306,Jul. 2001.
[FAB95] B. Le Floch, M. Alard, and C. Berrou, Coded orthogonal frequency division multiplex
[TV broadcasting], Proceedings of the IEEE, vol. 83,no. 6, pp. 982996, Jun. 1995.
[FB11] B. Farhang-Boroujeny, OFDM Versus Filter Bank Multicarrier, IEEE Signal Processing
Magazine, vol. 28, pp. 92-112, May 2011
[Fet14] G. Fettweis, The Tactile Internet Applications & Challenges, IEEE Vehicular
Technology Magazine, Mar. 2014, accepted for publication.
[GAS13] Ivan Gaspar, Nicola Michailow, Ainoa Navarro, Eckhard Ohlmer, Stefan Krone, and
Gerhard Fettweis. Low complexity gfdm receiver based on sparse frequency domain
processing. In Vehicular Technology Conference (VTC Spring), 2013 IEEE 77th, pages
16. IEEE, 2013.
[GMC+13] I. Gaspar, N. Michailow, A. Navarro Caldevilla, E. Ohlmer, S. Krone, and G. Fettweis,
Low Complexity GFDM Receiver Based On Sparse Frequency Domain Processing, in
Proceedings of the 77th IEEE Vehicular Technology Conference (VTC Spring13),
Dresden, Germany, Jun. 2013, pp. 16.
[GMF+14] Ivan Gaspar, Luciano Mendes, Maximilian Matthe, Nicola Michailow, Andreas Festag,
and Gerhard Fettweis. Lte-compatible 5g phy based on generalized frequency
division multiplexing. In 1th International Symposium on Wireless Communication
Systems (ISWCS), 2014.
[GMM+14] Ivan S Gaspar, Luciano L Mendes, Nicola Michailow, and Gerhard Fettweis. A
synchronization technique for generalized frequency division multiplexing. EURASIP
Journal on Advances in Signal Processing, 2014(1):67, 2014.
[HL99] Hoeher, P.; Lodge, J.; Turbo DPSK: Iterative Differential PSK Demodulation and
Channel Decoding, IEEE Trans. Commun., Vol. 47, No. 6, pp. 837843, Jun. 1999.
[HT99] Hoeher, P.; Tufvesson, F.; Channel Estimation with Superimposed Pilot Sequence,
in Proc. of IEEE Global Telecommun. Conf. (Globecom99), Vol. 4, pp. 21622166,
Dec. 1999.
[JW07] P. Jung and G. Wunder, The WSSUS Pulse Design Problem in Multicarrier
Transmission, IEEE Trans. on Communications, 2007.
[KM98] W. Kozek and A. Molisch, Nonorthogonal pulseshapes for multicarrier
communications in doubly dispersive channels, IEEE Journal Sel. Areas in Commun.,
vol. 16, no. 8, pp. 15791589, 1998.
[KWJ+14] M. Kasparick, G. Wunder, P. Jung, D. Maryopi, Bi-Orthogonal Waveforms for 5G
Random Access with Short Message Support, European Wireless Conference,
Barcelona, Spain, May 14-16, 2014.
[WKJ15] G. Wunder, M. Kasparick, P. Jung, Interference Analysis for 5G Random Access with
Short Message Support , European Wireless Conference, Budapest, Hungary, 2015,
to appear.
FP7 ICT GA 318555
IR2.1
[WGA15] G. Wunder, S. A. Gorgani, S. S. Ahmed, Waveform Optimization Using Trapezoidal

Pulses for 5G Random Access with Short Message Support , SPAWC2015,
Stockholm, Sweden, 2015, invited paper.
[WKJ+15] G. Wunder, M. Kasparick, P. Jung, Spline Waveforms and Interference Analysis for
5G Random Access with Short Message Support, http://arxiv.org/abs/1501.02917,
submitted
[Lau09] R. S. Laugesen, Gabor dual spline windows, Applied and Computational Harmonic
Analysis, vol. 27, no. 2, pp. 180194, Sep. 2009.
[LW00] K. Lee and D. Williams, A space-time coded transmitter diversity technique for
frequency selective fading channels, in Proceedings of the IEEE Sensor Array and
Multichannel Signal Processing Workshop. Cambridge, MA: IEEE, Aug. 2000, pp. 149
152.
[Mat15] M. Matth, GFDM Transceiver using Precoded Data and Low-complex Multiplication
in Time Domain, Submitted to IEEE Communications Letters, 2015
[MB10] L. Mendes and R. Baldini, Performance of WHT-STC-OFDM in Mobile Frequency
Selective Channel, in Proc. The 7th International Telecommunications Symposium,
Manaus, Brazil, 2010.
[MF13] N. Michailow and G. Fettweis, Low Peak-to-Average Power Ratio for Next
Generation Cellular Systems with Generalized Frequency Division Multiplexing, in
Proceedings International Symposium on Intelligent Signal Processing and
Communications Systems (ISPACS13), Naha, Okinawa, Japan, Nov. 2013, pp. 651
655.
[MGK+12] Nicola Michailow, Ivan Gaspar, Stefan Krone, Michael Lentmaier, and Gerhard
Fettweis. Generalized frequency division multiplexing: Analysis of an alternative
multicarrier technique for next generation cellular systems. In Wireless
CommunicationSystems (ISWCS), 2012 International Symposium on, pages 171175.
IEEE, 2012.
[MKL+12] N. Michailow, S. Krone, M. Lentmaier, and G. Fettweis, Bit Error Rate Performance
of Generalized Frequency Division Multiplexing, in Proceedings 76th IEEE Vehicular
Technology Conference (VTC Fall12), Qubec City, Canada, Sep. 2012, pp. 15.
[MLG+15] Maximilian Matth, Luciano Leonel Mendes, Ivan Gaspar, Nicola Michailow, Dan
Zhang andGerhard Fettweis. Multi-User Time-Reversal STC-GFDM for Future Wireless
Networks. Under review EURASIP Signal Processing
[MLR+11] N. Michailow, M. Lentmaier, P. Rost, and G. Fettweis, Integration of a GFDM
Secondary System in an OFDM Primary System, in Proc. Future Network & Mobile
Summit, Warsow, Poland, Jun. 2011
[MMF14] Maximilian Matth, Luciano Leonel Mendes, and Gerhard Fettweis, Generalized
Frequency Division Multiplexing in a Gabor Transform Setting, IEEE Commun. Lett.,
accepted for publication
[MMF15] Maximilian Matth, Luciano Leonel Mendes, Gerhard Fettweis. Asynchronous Multi-
User Uplink Transmission with Generalized Frequency Division Multiplexing,
submitted for publication at ICC 2015
[MMI+14] Nicola Michailow, Maximilian Matth, Ivan Simes Gaspar, Ainoa Navarro Caldevilla,
Luciano Leonel Mendes, Andreas Festag, and Gerhard Fettweis. Generalized
frequency division multiplexing for 5th generation cellular networks. Transactions on
Communications, 2014.
[MP11] P. Marsch and G. Fettweis, Coordinated Multi-Point in Mobile Communications -
From Theory to Practice, Cambridge University Press, Aug. 2011.
FP7 ICT GA 318555
IR2.1
[NGV11] D. Noguet, M. Gauthier, and V. Berg, Advances in Opportunistic Radio Technologies

for TVWS, EURASIP Journal on Wireless Communications and Networking, Nov.
2011, http://jwcn.eurasipjournals.com/content/pdf/1687-1499-2011-170.pdf
[PLK06] Ping, Li; Liu, Lihai; Wu, Keying; Leung, W.K.; Interleave-Division Multiple-Access,
IEEE Trans. Wireless Commun., Vol. 5, No. 4, pp. 938947, Apr. 2006.
[PK11] V. D. Papoutsis and S. A. Kotsopoulos, Efficient rate adaptive resource allocation
scheme in uplink OFDMA wireless systems, in ICWMC 2011 : The Seventh
International Conference on Wireless and Mobile Communications. Luxembourg:
IARIA, 2011.
[Pre99] V. D. Prete, Estimates, decay properties, and computation of the dual function for
Gabor frames, Journal of Fourier Analysis and Applications, 1999.
[RS97] A. Ron and Z. Shen, WeylHeisenberg frames and Riesz bases in L 2(Rd), Duke
Math. J., vol. 89, no. 2, pp. 237282, 1997.
[RSH11] F. Rezaei, M. Hempel, and H. Sharif, LTE PHY Performance Analysis Under 3GPP
Standards Parameters, in Proc. IEEE 16th International Workshop on Computer
Aided Modeling and Design of Communication Links and Networks (CAMAD). Kyoto,
Japan: IEEE, Jun. 2011, pp. 102106
[SMH02] D. Schafhuber, G. Matz, and F. Hlawatsch, Pulse-shaping OFDM/BFDM systems for
time-varying channels: ISI/ICI analysis, optimal pulse design, and efficient
implementation, in 13th IEEE International Symposium on Personal, Indoor and
Mobile Radio Communications, vol. 3, 2002, pp. 10121016 vol.3.
[SSS14] F. Schaich, B. Sayrac, M. Schubert, On the Need for a New Air Interface for 5G, IEEE
ComSoc Multimedia Communications Technical Committee E-Letter, Sept. 2014.
[STB09] S. Sesia, I. Toufik, and M. Baker, LTE, The UMTS Long Term Evolution: From Theory to
Practice. Wiley Publishing, 2009.
[SW] Schaich, F.; Wild, T.; UFMC OFDM Transceiver Chain,
http://www.5gnow.eu/node/63
[SW14] Schaich, F.; Wild, T.; Relaxed Synchronization Support of Universal Filtered Multi-
Carrier including Autonomous Timing Advance, in proceedings of IEEE ISWCS14,
Barcelona, August 2014
[SWC14] Schaich, F.; Wild, T.; Chen, Y.; Waveform contenders for 5G - suitability for short
packet and low latency transmissions, in proceedings of IEEE Veh. Technol. Conf.
Spring (VTC14 Spring), Seoul, Korea, May 2014.
[TB04] C. Tan and N. Beaulieu, Transmission Properties of Conjugate-Root Pulses, IEEE
Transactions on Communications, vol. 52, no. 4, pp. 553558, Apr. 2004.
[Ung82] G. Ungerboeck, Channel coding with mullti-level/phases signals, Information
Theory, IEEE Trans. on, vol. 28, no. 1, pp. 55-67, Jan. 1982.
[UUE+12] K. Ueda, T. Uozumi, R. Endo, T. Nakamura, T. Heima, and H. Sato, A Digital PLL with
Two-Step Closed-Locking for Multi-Mode/Multi-Band SAW-Less Transmitter, in
Proc. IEEE Custom Integrated Circuits Conference, San Jose, USA, Sep. 2012, pp. 14.
[VWS13] Vakilian, V.; Wild, T.; Schaich, F.; ten Brink, S.; Frigon, J.-F.; Universal Filtered Multi-
Carrier Technique for Wireless Systems Beyond LTE, 9th Int. Workshop on
Broadband Wireless Access at IEEE Globecom13, Atlanta, Dec. 2013.
[WG15] G. Wunder, S. Afrasiabi Gorgani, Overshooting and Lp norms of class of Nyquist
filters, to be published, March 2015
[WJW14] G. Wunder, P. Jung, C. Wang, Compressive Random Access for Post-LTE Systems,
IEEE ICC Workshop on Massive Uncoordinated Access Protocols, Sydney, Australia,
June 10-14, 2014
[WJK+14] Wunder, G.; Jung, P.; Kasparick, M.; Wild, T.; Schaich, F.; Chen, Y.; Brink, S.T.; Gaspar,
I.; Michailow, N.; Festag, A.; Mendes, L.; Cassiau, N.; Ktenas, D.; Dryjanski, M.;
Pietrzyk, S.; Eged, B.; Vago, P.; Wiedmann, F., 5GNOW: non-orthogonal,
FP7 ICT GA 318555
IR2.1
asynchronous waveforms for future mobile applications, Communications

Magazine, IEEE , vol.52, no.2, pp.97,105, February 2014
[WKJ15] G. Wunder, M. Kasparick, P. Jung, Spline Waveforms and Interference Analyis for 5G
Random Access with Short Message Support, Arxiv e-prints, Jan. 2015
[WSC14] Wild, T.; Schaich, F.; Chen, Yejian; 5G Air Interface Design based on Universal
Filtered (UF-)OFDM, in Proc. 19th Int. Conf. Digital Signal Process. (DSP14), Aug.
2014.
[WWS+14] Wang, X.; Wild, T.; Schaich, F.; Santos, A., Universal Filtered Multi-Carrier with
Leakage-Based Filter Optimization, European Wireless, EW14, Barcelona, May 2014
[GW15] S.A. Gorgani and G. Wunder, Derandomized Multi-block Sign Selection for PMEPR
Reduction of FBMC Waveform, VTC 2015 Fantastic 5G Workshop on 5G New Air
Interface, to appear
--- End of Document Main Body---
FP7 ICT GA 318555

5GNOW D3.3 Final

Uploaded by

Copyright:

Available Formats

5GNOW D3.3 Final

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

5GNOW D3.3 Final

Uploaded by

Copyright:

Available Formats

The 5GNOW Project Consortium groups the following organizations:

Partner Name Short name Country

Final 5GNOW Transceiver and frame structure concept

The 5GNOW Project Consortium groups the following organizations:

Partner Name Short name Country

Partner Author name

Title: Final 5GNOW Transceiver and frame structure concept

Last Update: Monday, May 04, 2015

No. Version Edition Author(s) Date

1.1 Sporadic traffic

1.2 Spectral and temporal fragmentation

1.3 Real-time constraints

Summarizing, although OFDM could be tuned to address different granularity of scheduling

2.1 Waveform candidates

In 5GNOW the following waveform approaches have been developed:

Generalized Frequency Division Multiplexing (GFDM) is an innovative multicarrier modulation

2.2.1 System Description and Properties

Figure 2.2.1: Block diagram of the transceiver

Collecting the filter samples in a vector g k , m g k , m [ n ] allows to formulate (2.2.1.2) as

where A is a KM KM transmitter matrix [MKL+12] with a structure according to

Transmission through a wireless channel is modelled by y H x w , where y is the received

transceiver equation can be written as y H A d w . Introducing z H 1 H A d H 1 w A d w

where B is a KM KM receiver matrix.

(a) A transmitter matrix (b) Three columns in detail

which are then passed to a decoder to obtain b .

(a) OFDM (b) SC-FDE (c) SC-FDM and GFDM

2.2.2 Fast computation of the receive filter

2.2.3 Multi-User Time-Reversal Space-Time Coding for GFDM

2.2.3.1 Time-Reversal Space-Time Coding

the transmitted blocks are acquired by inverse Fourier transform

2.2.3.2 Multi-User STC-GFDMA

Figure 2.2.4: Simplified block diagram of the TR-STC-GFDMA system

At the BS, the received signal in the frequency domain is given by

Universal filtered multicarrier (UFMC) can, roughly speaking, be interpreted as a generalization of

Figure 2.3.1 Basic UFMC transceiver chain

2.3.1 UFMC Frequency- and Time Domain Properties

UFMC, L = 80, SLA = 60

2.3.2 Channel estimation and equalization

2.3.3 Native MIMO/CoMP support

2.3.5 Adaptivity Potential

2.4.1 FBMC principles

s0 t pTx t e j 2 f0t pRx t e j 2 f0t s0 n

s1 t pTx t e j 2 f1t pRx t e j 2 f1t s1 n

sNc 1 t pTx t e j 2 f Nc1t pRx t e j 2 f Nc1t sNc 1 n

Figure 2.4.1: Block diagram of a multicarrier transceiver

Figure 2.4.2: FBMC - filters for subcarriers 0 (blue) and 1 (red), = , =

1 is the vector of data for the th FBMC in the time domain.

Figure 2.4.3: FBMC transmitter with filtering in the frequency domain

Figure 2.4.4: Frequency spreading

2.4.2 FBMC receiver

Figure 2.4.5: FS-FBMC receiver principles

2.5.1 Transmitter and receiver Structure

The BFDM-based transmitter structure is illustrated schematically in Figure 2.5.1.

Figure 2.5.1 BFDM based pulse shaped PRACH transmitter

The BFDM-based PRACH receiver is illustrated schematically in Figure 2.5.2.

Figure 2.5.2 BFDM based pulse shaped PRACH receiver

2.5.2 Pulse design

2.5.2.1 Spline waveforms