Increasing handset performance using True polarization diversity

Valenzuela-Valdés, Juan F. Sánchez-Hernández, David A.

EMITE Ing Technical University of Cartagena
Edificio CEEIM. Campus Espinardo Plaza del Hospital, 1
E-30100 Murcia , SPAIN E-30202 Cartagena, SPAIN
[email protected] [email protected]

Abstract— While many different simulation tools are available to including delay and signaling constraints and the statistical
compute MIMO capacity and diversity gain system performance, nature of the channel. The instantaneous channel capacity for
few works tackle the problem from the measurements MIMO systems is well defined by,
standpoint. In this paper, several measurements of MIMO
capacity and diversity gain performance of dipole antenna arrays ⎛ ⎛ SNR ⎞⎞
are performed through the use of a reverberation chamber, C MIMO = log 2 ⎜⎜ det⎜ I R + * H * H ' ⎟ ⎟⎟ bits/s/H z (1)
which includes many effects present in real MIMO channels but ⎝ ⎝ T ⎠⎠
generally avoided in simulations. Results show that polarization
The system has T antennas at the transmitter and R antennas
diversity can be effectively combined to spatial diversity even for
Rayleig-fading MIMO scenarios to achieve increased diversity at the receiver and IR is the identity matrix with dimension R.
gain and MIMO capacity. In an independent identically distributed (i.i.d.) Rayleigh
environment this capacity can be approximated for high
Index Terms— Diversity gain, MIMO, spatial diversity, signal-to-noise ratio (SNR) to C≈min(T,R)log2(SNR) bits/s/Hz.
polarization diversity, reverberation chamber. This growth potential is extraordinary since each 3-dB
I. INTRODUCTION increase in transmit SNR will result in roughly min(T,R)
b/s/Hz capacity increase in comparison to 1 b/s/Hz capacity
The improvement granted by polarization diversity in wireless
gain in single-antenna systems. For correlated Rayleigh-fading
systems is typically obtained by an additional de-correlated
environments, however, which represent real scenarios due to
channel provided by a polarization state made orthogonal to
lack of scattering at different antenna elements at reception
the existing one, usually at the transmitting end. A randomly
oriented linearly-polarized antenna is also typically used at the and/or transmission ends, the capacity slope changes as a
receiver for evaluating polarization diversity. In this scheme function of the number of antennas in reception and
the cross-polarization discrimination (XPD) factor is the usual transmission, among other parameters, and more detailed
evaluation parameter, with low correlation coefficients being simulation studies are required and available in the literature
achieved in LOS and non-LOS situations [1]. Some [4].
combinations of two-branch orthogonal polarization and The multipath environment can be generated artificially in a
spatial diversity have been reported [2]. In mobile reverberation chamber fed by wall-mounted antennas, which
communications scenarios, however, multiple scattering may thereby provides a statistically repeatable laboratory-produced
not be sufficient for a given polarization to decouple half its environment for characterizing mobile terminals and their
power into the orthogonal polarization [3]. A recent letter [4] antennas. A reverberation chamber is a metal cavity
has proposed a novel true polarization diversity (TPD) sufficiently large to support many resonant modes (multimode
technique by rotating one antenna by a certain angle with cavity), which are perturbed with stirrers in order to create a
respect to the contiguous element in the MIMO array. In this fading environment similar to the ones found in indoor and
way an arbitrary angular separation between contiguous urban environments but with a uniform elevation distribution
dipoles is employed in an equivalent way that an arbitrary of the incoming waves. The reverberation chamber can be
spatial separation is employed for spatial diversity. In efficiently used to measure radiation efficiency, which
addition, an accurate prediction of the correlation coefficient characterizes the performance of single antennas in an
between two dipoles separated by both a spatial distance and
isotropic multipath environment according to the above. The
an arbitrary angular position has not been available until very
received signal becomes then normally (Gaussian) distributed,
recently [5]. This has made possible the evaluation of TPD
and from this their associated magnitudes get a Rayleigh
performance for handset MIMO. Results to be presented in
this contribution demonstrate that under Rayleigh-fading distribution, and the phases get a uniform distribution over 2π.
scenarios TPD can be effectively combined with spatial The system is completed with an elevation distribution factor
diversity to nearly double the diversity gain and MIMO to account for the larger probability of having waves coming
capacity for the same available volume. from close to vertical than from close to horizontal directions
in real multipath environments and with the use of the
II. MEASUREMENT TECHNIQUE AND SET-UP embedded element pattern as a way to emulate the fact that for
The capacity for fading channels can be defined in a real multipath environments the received signal on each port is
transmitted or detected independent of other ports. The
number of ways, very much depending upon the amount of
channel knowledge made available to the transmitter, chamber can also be used to measure the diversity gain and
channel capacities of MIMO antennas with useful and

978-1-4244-2517-4/09/$20.00 ©2009 IEEE 1

Authorized licensed use limited to: BEIHANG UNIVERSITY. Downloaded on June 03,2021 at 02:46:38 UTC from IEEE Xplore. Restrictions apply.
accurate repeatability. Since all excitations and parameters can probability level. Channel capacity, also known as spectral
be weighted, different fading scenarios can also be obtained. efficiency is calculated using the channel estimates hmn in eqn.
With the use of a reference antenna and the MIMO antenna (5) between each of the MIMO array antenna and each one of
array several MIMO parameters can be evaluated through the wall-mounted exciting antennas, i.e., a m x n radio
adequate processing of the measured S-parameters, which are channels represent a m x n MIMO system, with n being the
gathered between the measured port and the wall-mounted number of MIMO array antennas under test and m the number
antennas for all positions of the platform and mechanical of wall-mounted antennas. With only one wall-mounted
stirrers and for all frequency points. The measurement antenna the normalized channel estimates form the channel
procedure is then repeated for every antenna port, with the vector becomes,
uncorrected ports terminated in 50 Ω, for exactly the same
stirrer positions and position of the array inside the chamber,
H 1 x n = [h11 h12 ... h1n ] (6)
and a similar procedure is employed with the reference For the maximal ratio combining, for instance, the channel
antenna. Thus, the field environment is exactly the same when capacity C1 x n for the system can easily be calculated as,
measuring every port.
2
C1 x n = log 2 (1 + SNR∑i =1 h1i )
S-parameters are first pre-processed on the wall-mounted n
(7)
exciting antennas (Smn) and of the MIMO array antennas (Snn)
with a complex averaging over stirrer positions by, For each channel matrix estimates H1 x n, the channel capacity
1 1 is calculated for a specific SNR range, and all channel capacity
S mn =
N
∑ stirrerpos
S mn and S nn =
N
∑ stirrerpos
S nn (2) estimates are averaged to produce a maximum average
channel capacity as a function of the SNR. Maximal ratio
Where N is the total number of stirrer positions. In fact, combining represents the maximum theoretical capacity and
averaging is also performed for platform positions and the upper boundary. Since coupling is inherently accounted for
polarization stirring at the user’s perusal. At low frequencies it in the measurements, mean capacity is then derived by,
is also advantageous to perform averaging over a small
frequency band (frequency stirring) to get more independent SNR
C m x n = log 2 (det( I R + H m x n H m* x n )) (8)
field samples representing a richer multipath environment. m
Since some commercial systems like GSM use frequency Similarly, other MIMO parameters such as the correlation
hoping, this is yet another more realistic emulating coefficients can also be evaluated in the chamber.
characteristic of the reverberation chamber. S-parameters are
then normalized to, III. MEASUREMENT SCENARIOS
Smn Reverberation chambers have already demonstrated their
′ =
Smn (3) ability to reproduce multipath propagation environments
2 2
1 − ( Smn ) 1 − ( S nn ) typically found in indoor and urban wireless environments [6].
to obtain a better accuracy for the radiation efficiency. The In order to evaluate the full potential of TPD techniques,
average net transfer function of the chamber becomes, combined-diversity systems with both spatial and TPD
techniques have been tested. Measurements were performed for
1 different 3x6 linear MIMO systems. The diverse scenarios are
∑
2
T= stirrerpos
′
S mn (4) illustrated in figure 1. The correlation coefficients and MIMO
N capacity performance were measured for the MIMO array
S-parameters can also be weighted with mismatch factors formed by the 3 wall-mounted transmission antennas and the
before the frequency stirring, and corrected to a reference combination of 6 receiving dipole antennas. The only
level corresponding to 100 % radiation efficiency. The possibility of using conventional orthogonal polarization
processed S-parameters represent estimates of the channel diversity (OPD) in this scheme is to alternate the polarization
matrix H of multipath communication channels set up between orthogonal sate between contiguous elements (VHVHVH).
the wall antennas and the MIMO array inside the chamber. Such a system is really a particular case of TPD with dθ=90º.
Apparent diversity gain for the selection- or maximal ratio In fact, the spatial-only diversity scheme can also be
combining techniques is obtained from the processed S- considered a particular case of TPD with dθ=0º.
parameters by evaluating the cumulative probability
distributions of the measured channel samples received at each
MIMO array antenna by,
2
Smn eref (1 − Snn )
hmn = (5)
Tref
Figure 1. Tested TPD array schemes
where Tref is the net chamber transfer function for the
reference antenna and eref is its known radiation efficiency.
The effective diversity gain is the increase from the reference
level to the combined signal that is observed at the 1%

2
Authorized licensed use limited to: BEIHANG UNIVERSITY. Downloaded on June 03,2021 at 02:46:38 UTC from IEEE Xplore. Restrictions apply.
 IV. MEASUREMENT RESULTS
Measured correlation coefficients for the linear arrays are
illustrated in figure 2 for small element spacing. The
correlation coefficients depicted in figure 2 are measured with
respect to the first dipole in the array. The results confirmed
the enormous potential of TPD for combined-diversity
schemes. It is easily observed from figure 2 that the alternating
orthogonal polarization scheme (TPD with dθ=90º) has a
jigsaw correlation behaviour, which could diminish MIMO
capacity. It is also interesting to observe from figure 2 that
more general TPD schemes (different from dθ=90º) depict a
different correlation pattern with respect to more conventional
OPD. Diversity gain measurements were performed in the
reverberation chamber for two receiving dipoles separated by
both an angular and a spatial separation. Figure 3 shows how
the measured diversity gain depend on angular separation dθ
(º) and wavelength-normalized dipole spatial separation
D=(d/λ). As expected, the combination of both spatial
Figure 3. Measured diversity gain (dB) vs. angular separation dθ (º)
diversity and TPD provided increased diversity gain with only
with spatial separation D=(d/λ) as a parameter for 3x2 MIMO
two elements in the array. The combination has a stronger systems.
effect when both separations are not large, i.e. when the spatial
separation is large (D ≥ 0.24), the angular separation can
hardly improve the diversity gain, and vice versa when the
angular separation is large (dθ ≥ 54º), the spatial separation
can barely improve the diversity gain. This suggests that a
good combination of the two techniques represents the most
efficient technique for optimum diversity performance within
the same reduced volume made available to the complete
array. This is also expected to have an effect on MIMO
capacity. Figure 4 depicts the measured MIMO capacities for
different combined-diversity systems at SNR=15 dB. All
combined-diversity tested systems provide increased capacity
with respect to the spatial-diversity-only linear MIMO system.
In addition, when the spatial-diversity antenna spacing D is
small enough (D<0.07) to provide correlation coefficients
over 0.5, a considerable capacity reduction is observed for
spatial-diversity-only systems in comparison to combined-
diversity systems. It is also interesting to observe from the Figure 4. Measured 3x6 MIMO capacity for diverse scenarios.
previous figure that a combination of spatial and TPD
techniques performs at nearly full capacity, even for extremely
small spatial antenna spacing (D=0.01). At SNR=15 dB and
for D<0.02, the combination of spatial and TPD techniques
nearly doubles the MIMO capacity of the spatial-only linear V. CONCLUSIONS
MIMO system. While spatial diversity has already been identified in the
literature as an excellent candidate for improving diversity
gain and MIMO capacity for wireless systems, results
presented in this paper demonstrate that true polarization
diversity (TPD) is equally important and particularly
significant when combined to spatial-diversity schemes. TPD
can be effectively combined with spatial diversity to nearly
double MIMO capacity for the same available volume, which
could be useful for handset MIMO.

VI. REFERENCES
[1] J.J.A. Lempiainen and J.K. Laiho-Steffens, “The Performance of
Polarization Diversity Schemes at a Base Station in Small/Micro Cells at
1800 MHz,” IEEE Transactions on Vehicular Technology, Vol. 47, No.
3, pp. 1087-1092, August 1998.
[2] K. Cho, T. Hori and K. Kagoshima, “Effectiveness of Four-Branch
Figure 2. Measured envelope correlation coefficients for D=0.02. Height and Polarization Diversity Configuration for Street Microcell,”

3
Authorized licensed use limited to: BEIHANG UNIVERSITY. Downloaded on June 03,2021 at 02:46:38 UTC from IEEE Xplore. Restrictions apply.
 IEEE Transactions on Antennas and Propagation, Vol. 6, No. 6, pp. 776- [5] J.F. Valenzuela-Valdés, A.M. Martínez-González, and D. Sánchez-
781, June 1998. Hernández, “Estimating combined correlation functions for dipoles in
[3] M. Kang and M-S. Alouni, “Capacity of correlated MIMO Rayleigh Rayleigh-fading scenarios,” IEEE Antennas and Wireless Propagation
Channels,” IEEE Transactions on Wireless Communications, Vol. 5, No. Letters, Vol. 6, pp. 349-352, 2007.
1, pp. 143-155, Jan. 2006. [6] K. Rosengren and P.S Kildal, “Study of distributions of modes and plane
[4] J.F. Valenzuela-Valdés, M.A. García-Fernández,. A.M. Martínez- waves in reverberation chamber for characterizacion of antenas in
González, and D. Sánchez-Hernández,, “The role of polarization multipath environment,” Microwave and Optical Technology Letters,
diversity for MIMO systems under Rayleigh-fading scenarios,” IEEE Vol. 30, pp. 386-391, 2001.
Antennas and Wireless Propagation Letters, Vol. 5, pp. 534-536, 2006.

4
Authorized licensed use limited to: BEIHANG UNIVERSITY. Downloaded on June 03,2021 at 02:46:38 UTC from IEEE Xplore. Restrictions apply.