Software-Defined Radio

SDR platform enables reconfigurable direction

finding system
A software-defined radio (SDR) transceiver platform is used to implement a
reconfigurable direction finding (DF) system. This article will describe the
intended data flow and data rate requirements of a typical DF system, map the
DF software components to the SDR platform, and demonstrate the availability
of resources for more complex implementations.
By Jyothsna Rajan

D irection finding (DF) is the process of

obtaining the angle of arrival bearings
of radio signal emitters. DF has multiple
system or tracing a signal interference source
or locating a hostile target.
Most DF methods inherently use the
nae, arranged in a known configuration, to
calculate its bearing.
A typical DF system must allow the user
military and commercial applications, such phase of the incident signal wave front and to perform spectral analysis over a predefined
as remotely guiding an aircraft navigation its relative time-of-arrival at multiple anten- frequency band, select a signal of interest

Figure1. Direction finding demonstration system processing blocks.

36 http://www.rfdesign.com January 2005

 ■ channelize the data to extract the signal/s
of interest:
■ provide high-speed data transfer capabil-
ity; and
■ simultaneously process channel data from
multiple receivers to deduce the angle of
arrival (AOA).
Although a variety of available processing
boards and I/O modules can cater to one
or more of the above functions, there are
few systems that can provide all of the
functionality in a seamlessly integrated
environment. One system that is capable of
combining the above-mentioned fundamental
functions is Spectrum’s flexComm™
SDR-3000 integrated development system.
This commercial off-the-shelf (COTS)
platform was recently used to demonstrate
a software-defined DF system. Using this
example, this paper will discuss the architec-
ture of the SDR-3000 system, describe the
intended data flow and data rate require-
Figure 2. Software task mapping. ments of a typical DF system, map the
Figure 2. Software task mapping. DF software components to the SDR-3000
hardware components and demonstrate the
and then locate the source. This requires a characteristic variations, environmental availability of resources for more complex
system that can: effects, etc.; implementations.
■ digitize received analog IF signals; ■ perform windowing and digital filtering;
■ synchronize data from multiple receivers; ■ execute FFT algorithms on the sampled Why SDR-3000?
■ calibrate data to account for component data; The SDR-3000 provides a hardware

Circle 29 or visit freeproductinfo.net/rfd

38 http://www.rfdesign.com January 2005

 PowerPC™; and
■ baseband processing board (PRO-3500)
with two onboard Motorola MPC7410
PowerPCs and an IBM 405GP PowerPC.
The system provides the combined high-
speed logic execution capability of the
Virtex-II FPGAs with the algorithm execu-
tion capability of PowerPC MPC7410
G4 processors. A multi-input transition
module (TM1-3xxx) enables the system
to simultaneously receive inputs from
different receivers, digitize and then transmit
the data to the signal conditioning board
(PRO-3100) via a high-speed digital IF
fabric at rates up to 1.28 Gbps. Spectrum’s
flexFabric communications architecture,
based on RapidIO™ technology, facilitates
interprocessor communications throughout
the system for high-speed deterministic
data flows, while PCI-based communications
is used for “soft real-time” payload data
and control. In addition, the PRO-3100
and the PRO-3500 boards are network
Figure 3. GUI display. enabled, allowing the SDR-3000 to easily
operate as part of a larger system.
architecture and system configuration that ■ multi-input transition module (TM1-3xxx);
enables users to combine all of the ■ signal conditioning board (PRO-3100) Application framework
fundamental functions required in a DF with four user-programmable Xilinx® A basic application was developed to
application. The standard SDR-3000 system Virtex-II™ XC2V6000 field-programmable demonstrate the capability of the SDR-3000
consists of: gate arrays (FPGAs) and an IBM 405GP to operate independently or as part of a

Circle 31 or visit freeproductinfo.net/rfd

40 http://www.rfdesign.com January 2005

 bins is displayed. The zoom feature on
the GUI allows the user to increase the
bin resolution on the display. During opera-
tion, the user selects a frequency of
interest from the base spectral display, and
continually zooms in until the display resolu-
tion is one bin per pixel.
Using the mouse, the user can then
select the signal of interest, and the GUI
transmits the pixel index of the selected SOI
to a processor that uses the zoom-level
and the display width to identify the bin
Figure 4. DFDS(S) system configuration. number. A bin extractor, executing on the
FPGA, extracts the in-phase (I) and quadra-
larger DF system. This application enabled of channel data storage memory to 64 Mbytes/ ture (Q) data corresponding to the
the user to: input mandates that the signal of interest required bin for each of the four inputs
■ receive and process inputs coherently be on air for a minimum of 2.5 seconds and forwards it to an MPC7410 on the
from four high-frequency (HF) antennae in order to be detected and its AOA calcu- PRO-3500. This processor first calculates
channels; lated. The same channel data is used to the phase for each input and then calculates
■ calibrate the input channels; calculate the power spectral density (PSD) the AOA of the SOI from the phase data
■ display a composite power spectrum on for each channel. Since the aim of this for the individual inputs. This AOA value
a host computer; first stage is merely to identify a strong is then transmitted to the GUI for display
■ select a signal of interest (SOI) from a signal within the operating frequency range, on the compass rose.
spectral display; and the data corresponding to a channel is
■ calculate the AOA and display it on a summed across the four inputs and then Data flow in the basic DF
compass rose. displayed. Spectral averaging is used to application
To provide portability across OS and provide data at a rate sufficient to meet the Figure 1 shows the signal data path
operation in a distributed environment, a display update rates. and processing blocks implemented in the
CORBA-enabled client-server design was The GUI displays the PSD data for DF application. Two of the user-program-
used. A JAVA-based GUI acted as a client, the entire range (0 MHz to 40 MHZ). Since mable FPGAs are used to service sampled
and applications running on the MPC7410 the spectral display is only capable of data from four antennae and buffer the
processors acted as servers. displaying 800 to 1000 horizontal pixels, data for bin extraction in their local
The design emphasis was to demonstrate each pixel represents data from multiple SDRAMs. A bin extraction module within
the ability to use the SDR-3000 system FFT bins or frequency channels, with each FPGA calculates the bin number
as a direction finder and not to develop the highest PSD value within the set of corresponding to the selected SOI and
a sophisticated DF system in its entirety.
Accordingly, greater attention is paid
to implementing typical DF application
tasks and meeting functional requirements.

Concept of operation
The current design uses the IF sampling
conversion technique [1] using an ADC
operating at 80 MSPS. Signal amplifiers
and low-pass filter at the front end limit
the operational frequency range between
10 MHz and 40 MHz.
For this application, the system used
FFT cores from RF Engines Ltd. (RFEL),
a technology partner that specializes in
advanced FPGA IP cores. RFEL’s FFT
core, capable of servicing two simultaneous
inputs, is used to channelize the data
into 2.5 kHz wide channels. The digital
filters in the FFT core have a cut-off
frequency of 40 MHz and an 80%
passband distributed evenly about the center
frequency, giving an operating frequency
range of 4 MHz to 36 MHz. The software
and hardware characteristics modify the net
frequency range to 10 MHz to 32 MHz.
Sixty four MB of SDRAM memory is used
to store channel data for about 2.5 seconds
for each of the four inputs. The restriction Figure 5. DFDS(T) system configuration.

42 http://www.rfdesign.com January 2005

 Table 1.
User-programmable FPGA 1 (processes the spectral data)
Device utilization summary: need for unnecessary high-speed bulk data
Number of External IOBs 184 out of 824 22% transfers without compromising system
Number of LOCed External IOBs 184 out of 184 100% performance."
Number of MULT18X18s 4 out of 144 2%
DF and calibration algorithms
Number of RAMB16s 45 out of 144 31% The DFCP incorporates a DF algorithm
Number of SLICEs 2862 out of 33792 8% suitable for use with fixed antenna systems.
Number of BUFGMUXs 8 out of 16 50% It uses the phase of the incident signal at
Number of DCMs 4 out of 12 33%
an antenna to detect the AOA. This simple
algorithm calibrates the DF data but
User-programmable FPGA 0 and 2 (implements the decimation/FFT core) does not compensate for multipath effects.
Device utilization summary: The DFCP also has the ability to average
Number of External IOBs 203 out of 824 24% a user-defined number of sequential DF
measurements before passing it on to
Number of LOCed External IOBs 203 out of 203 100%
the host.
Number of MULT18X18s 84 out of 144 58% The data from the four antennae traverse
Number of RAMB16s 136 out of 144 94% through different cables and hardware
Number of SLICEs 14787 out of 33792 43% components (filters, amplifiers, ADC/DAC)
prior to the DDC stage. The system must
be calibrated to compensate for relative
extracts the FFT data for that bin from The DF/calibration controller (DFCC) phase and amplitude imbalance caused by
the delay buffer. A third user-programmable and spectral data controller (SDC) are physical and electrical properties of the
FPGA is used to calculate the PSD for used to configure the FPGA application cables and hardware components. Calibra-
spectral analysis. A spectral averaging resources in response to commands from tion is mandatory the first time the system is
module on this FPGA allows the user the GUI. The controller blocks communicate powered on and must be repeated whenever
to control the rate at which the PSD data with the servers over the network using there is a change in hardware, cables, antenna
is sent to the PPC7410 processing nodes. TCP/IP sockets. configuration or at extended time periods
On the PRO-3500, one PPC7410 node The processing blocks described earlier to account for component degradation due
is used to execute the DF and calibration can be implemented on any of the available to aging. The current design assumes that
algorithms while the second node is used 405GP or PPC7410 processing nodes. calibration and DF processing are mutually
to scale spectral data corresponding to Different placement schemes were evaluated, exclusive functions and consequently uses
the zoom level selected on the spectral each requiring different intertask data a different set of firmware images for the
display. communication methods. The primary metrics user-programmable FPGAs for performing
used for comparison included, but are off-line calibration. Each time the DF
Task classification and mapping not limited to, the following: algorithm is executed on a selected data
Figure 2 shows the major software ■ data throughput (high, low); bin, run-time calibration uses the calibration
task breakdown for the basic DF application, ■ frequency of data transmission (frequent, data to compensate for inherent differences
data flow, task to processor mapping continuous, intermittent on request); introduced by the different components in
and the protocols used for intertask ■ protocol or transport medium; and the input data paths.
communication. ■ scope (within same processor, different
The graphical user interface (GUI) ex- processor on same board, processor on a Design verification
ecutes on a Windows PC and displays different board). Two configurations, varying only in
the frequency spectrum and bearing of The mapping scheme shown in Figure 2 the way the received signals are fed to the
the received signal. The GUI is written provides judicious use of resources and SDR-3000 system, were used to verify the
in JAVA and operates as a CORBA-enabled allows implementation of data manipulation system design. The DFDS(S) configuration,
client interacting with the CORBA-enabled techniques that eliminate the need for shown in Figure 4, uses a single signal source
DF and spectral data servers on the unnecessary high-speed bulk data transfers but different cable lengths to simulate the
PPC7410 nodes. Figure 3 shows a typical without compromising system performance. phase difference between signal incidents
display on the GUI. Many different software partitioning and at the different antennae.
The DF and calibration processor (DFCP) mapping schemes were evaluated. Of DFDS(T) configuration, uses antennae
task reads the selected bins of interest from these, the current software task mapping arranged in a rectangular array as shown
the bin extraction module on the FPGA scheme, shown in Figure 2, provides the in Figure 5, to receive CB radio frequency
and calculates the AOA. maximum processor idle times and free band- signals emitted by a COTS CB radio. The
The DF and calibration server (DFCS) width on the interboard and intraboard four COTS HF omni antennae were mounted
services DF and calibration-related data communication bus. at optimum spacing for operation up to
requests and configuration commands from "Many different software partitioning and 32 MHz.
the GUI. The spectral data server (SDS) mapping schemes were evaluated. Of these, All functions and operational features
services spectral display-related data the current software task mapping scheme, of the DFDS(S) were satisfactorily tested
requests and configuration commands. shown in Figure 2, provides the maximum in the presence of the end user. Cables
The spectral data processor (SDP) con- processor idle times and free bandwidth on of pre-calculated lengths were used to
tinuously fetches spectral data from the the interboard and intraboard communica- simulate signal source at cardinal points
user-programmable FPGAs and scales it tion bus through judicious use of resources. (N, S, E, W) and at ±45 from them (NE,
for display using the most recent zoom index This allows implementation of data NW, SE, SW). Testing covered the
and display area pixel count. manipulationtechniques that eliminate the entire operating frequency range from

44 http://www.rfdesign.com January 2005

 10 MHz to 32 MHz. The simulated system processing power, speed, and design and
displayed AOA to an accuracy of ±5. operational flexibility required to perform
DFDS(T) was tested in a parking the complex functions. Furthermore, the
lot setting exhibiting a minimal level of team was able to design, develop, procure
multipath. The spectral display successfully equipment, and test and validate the
showed peaks at frequencies equal to DF system on this platform in only four
that of received signals. The AOA display months.
for this case exhibited ±25 accuracy. The system can be expanded to accommo-
This ambiguity was expected, as the test date more input channels by adding addi-
environment was not completely multipath- tional transition module and PRO-3100
free and reflection-free while the implemented pairs. Channel data must be synchronized
DF algorithm assumed such ideal across all the transition modules. To
conditions. handle the increased data volume and
channelization requirements, additional
Capability demonstration and FPGA cores and software tasks may need
system acceptance to be incorporated and the software task
All system requirements were either mapping either re-partitioned or new soft-
met or exceeded. Despite using a simple ware tasks added.
DF algorithm, the DFDS(T) was able to This platform can also be used to
display AOA to within ±5 when tested efficiently perform multiple simultaneous
out in the open. System resource usage DFs on multiple signals of interest. This
statistics clearly depicted that sufficient may require the use of different channelization
resources were available to expand the techniques and probably, implementation of
system to include more inputs or use wideband and narrowband channelization.
more complicated algorithms. The delivered As an example, a user could implement
system, configured as a stand-alone unit, a beam former to scan a wide frequency
could be easily transported, set up and range, automatically detect signals with power
operated. above a preset threshold, and then use
Resource usage statistics for the narrowband channelization to extract the
user-programmable FPGAs are shown in signal of interest. RFD
Table 1.
The majority of user-programmable References
FPGA 0 and 2 slices are used up by 1. L. Pucker, “Channelization Techniques
RFEL’s FFT core with very little usage for Software-Defined Radio,” Proceedings
by application processing code. Most of of the SDR Forum Technical Conference,
FPGA 1 is still unused. This leaves an November 2003.
entire user-programmable FPGA and
associated SDRAM for any additional pro-
cessing and delay buffer implementation.
Depending on the size of the FFT/DDC
core used, more optimal designs can be ABOUT THE AUTHOR
developed to implement additional process- Jyothsna Rajan is an Applications
ing on these FPGAs, such as IFFT, mod/ Engineer with Spectrum Signal
demod or decoding. Processing Inc. in Burnaby, B.C.,
In the current design, the 7410s continu- Canada. With more than 13 years
ally extracts bin data and spectral data of experience, her areas of expertise
from the PRO-3100. With this processing, include digital signal processing and
the 7410 used for DF and calibration is lightly real-time embedded systems software
loaded while the 7410 used for spectral for distributed communications and
data processing is occupied most of the control. She received her Bachelor
time. However, with minor software modifi- of Engineering from Karnataka
cations that allow the two processors to Regional Engineering College in
process data only upon request from the India and a Masters in Electrical
user, the processor idle time can be consider- Engineering, specializing in DSP/
ably increased. communications from the University
of Saskatchewan in Saskatoon,
Conclusion Canada. Prior to joining Spectrum,
A direction finding system demands she worked under the aegis of
a high degree of functionality. Today, Telecommunications Research Lab
COTS platforms are sophisticated enough at Saskatoon developing a portable
for DF applications as successfully demon- wideband microseismic system for
strated with the flexComm SDR-3000. fault location in underground mines.
The platform provided the necessary

Circle 35 or visit freeproductinfo.net/rfd

46 http://www.rfdesign.com January 2005