|
Software Radio Research @ MPRG
Introduction
Software radio exploits Moore’s law, which states
that processors are ever increasing in computational speed while decreasing in
area, power, and cost, to perform the majority of a radio’s physical layer
processing in software, typically using microprocessors (mP)
and Digital Signal Processors (DSP) in a reconfigurable manner. Ideally, a
software radio would define and control all aspects of its physical and MAC
layer behavior in software, from its modulation to its center frequency and
bandwidth. However, this is not currently practical [1].
Thus while software radio conceptually is purely a
programming problem, in reality, software radio requires extensive research into
numerous seemingly unrelated areas. The need for a flexible RF front end has
spurred research into Micro Electro-Mechanical Systems (MEMS) [2]. The possibility of software radio has led to research into
artificial intelligence in an attempt to make cognitive radios, radios which are
able to learn and adapt to changes in their environment. The availability of
software control over hardware has led to re-examinations of the way radios are
designed such as digital predistortion for power amplifiers [3]. The potential impact that software radios can have on
networks has raised regulatory issues that the FCC is examining [4] and caused others to propose the use of game theory to
understand their interactions [5]. Ultimately, software radio is a multi-disciplinary area of
research requiring both specialized knowledge and a broad systems-level
understanding of how its components interact. The following highlights some of
the areas MPRG is currently researching under the broad umbrella of software
radio.
Reconfigurable Processors:
Figure 1: Stallion Processor
While attractive for their flexibility, mPs
and DSPs have difficulty providing the computational power required for the
complex real-time processing demands of cutting-edge wireless standards.
Because of this, many radios, particularly handsets, still commonly use ASICs
(Application Specific Integrated Circuits). However, this severely limits the
flexibility of the device. Thus there appears to be a tradeoff between
processing efficiency and flexibility.
Recognizing
this problem, many have proposed the use of reconfigurable processors [6][7] – processors whose components are optimized for an
application, but can be reprogrammed to be optimized for another application.
Field Programmable Gate Arrays (FPGAs) are the best known of the reconfigurable
processors, but can require several milliseconds to completely reconfigure [8].
MPRG
has been taking a different approach, preferring an architecture known as Custom
Computing Machines (CCMs), which like FPGAs can be optimized for different
applications, but because of a coarser component granularity, CCMs can be
reprogrammed in microseconds
[9]. During the GloMo project, MPRG demonstrated that a CCM
developed by the Reconfigurable Lab at Virginia Tech, the Stallion CCM shown in
Figure 1 [10], was suitable for software radio applications. Through a
project sponsored by Samsung, MPRG is currently studying ways to improve the
design of CCMs for wireless applications.
Software
Architectures:
Figure 2: SCA Component Diagram
Unlike in the PC world where a de-facto software
architecture emerged, under the leadership of US military [11] and the SDR Forum [12], the wireless world has collaborated to produce a
standardized software architecture for software radios known as the Software
Communications Architecture (SCA) pictured in Figure 2 [11]. The SCA is designed to ensure interoperability of software
and hardware between domains. A domain can be thought of as a logical grouping
of wireless devices in terms of capabilities and target application. For
instance, all mobile phones would be in one domain and all base stations would
be in another.
While desirable from a development/maintenance
point of view, this interoperability requires significant processing resources
which are typically not available in the handheld domain. Through funding from
General Dynamics, MPRG is currently examining ways of reducing the footprint of
the SCA while preserving the core functionalities. Eventually, this research
will produce a functional “SCA-lite” for handheld devices.
Smart
Antennas:
Figure 3: VT-Star Receiver
Software radio also provides a natural platform
for implementing flexible processing intensive algorithms like those used to
implement smart antennas. Smart antennas are an array of two or more antennas
whose signals are intelligently combined to achieve significant improvements in
signal quality. Like human audio processing, smart antennas can provide
significant signal selectivity by exploiting spatial differences, differences in
signal patterns, or differences in signal amplitudes. Smart antennas, however,
can be used at the receiver and the transmitter.
MPRG
has a long history of involvement in smart antenna research, in terms of
algorithm development [13], simulation [14], and experimentation [15]. For the experiments, MPRG makes use of both donated
equipment, like the SDR-3000 from Spectrum Signal Processing, and in house
equipment such as the VT-STAR system shown in Figure 3. MPRG has examined the
gains seen over a single link [16], and the capacity gains seen when implemented in a network
[17]. Through funding from
Samsung,
Motorola, the
Navy (NAVCITI), and Texas Instruments (TI), MPRG is looking at various aspects of antenna array
design and implementation at the base station and at the handset.
Network
Analysis:
As software radio becomes a reality, many have
suggested utilizing the reconfigurable nature of software radios to implement
real-time parameter adaptations in response to changing conditions [18] [19]. Over a single link, these adaptive protocols can be shown
to provide significant performance improvements, but within the context of a
network, these protocols become difficult to analyze with traditional
techniques. Through our work [20][21][22] and the work of others [5][23][24], game theory has been shown to be well suited for tackling
these kinds of problems. MPRG is currently examining the application of game
theory to adaptive wireless networks enabled by software radio technology. This
work is funded by the
Office of Naval Research,
Motorola, and
an
IREAN
fellowship.
Software Radio
Platforms:
MPRG performs a significant portion of its
software radio research on cutting-edge software radio platforms. Through a
donation from Spectrum Signal Processing, MPRG uses the SDR-3000, a four-channel
SCA-compliant transceiver to perform experiments on the SCA, smart antennas, and
the role of reconfigurable processing in the SCA. Through another donation from
Signia, MPRG used the Sunrise radio to implement adaptive modulation schemes [25] and is currently exploring direction finding techniques.
References:
[1] Reed, Jeffrey H. Software
Radio: A Modern Approach to Radio Engineering, Prentice Hall, 200, p 510.
[2] William H. Weedon, William J.
Payne, and Gabriel M. Rebeiz, “MEMS-Switched Re-configurable Antennas,” in IEEE Antennas and Propagation Society
International Symposium, 2001.
[3] Nizamuddin, Muhammad Ali,
Predistortion for Nonlinear Power Amplifiers with Memory,” Master’s Thesis
Virginia Tech, December 2002.
[4] Cognitive Radio Technologies
Proceedings [Online]
Available
[5] A.B. MacKenzie, and S.B. Wicker.
“Game Theory in Communications: Motivation, Explanation, and Application to
Power Control,” Globecom 2001, pp. 821 -826.
[6] Brakensiek, J.; Oelkrug, B.;
Bucker, M.; Uffmann, D.; Droge, A.; Darianian, M.; Otte, M, "Software radio
approach for reconfigurable multi-standard radios," The 13th IEEE
International Symposium on Personal, Indoor and Mobile Radio Communications,
pp. 110 -114.
[7] Master, P., “The Age of Adaptive
Computing Is Here,” Field-Programmable Logic and Applications, 12th
International Conference, FPL 2002, Montpellier, France, September 2-4,
2002.
[8] Srikanteswara, Srikathyayani, James
Neel, Dr Jeffrey H. Reed, Dr. Peter M. Athanas “Soft Radio Implementations for
3G and Future High Data Rate Systems,” Globecom 2001, Volume: 6, pp.
25-29.
[9] Neel, James O., “Simulation of an
Implementation and Evaluation of the Layered Radio Architecture,” Master’s Thesis
Virginia Tech, December 2002.
[10] Soni, Maneesh, “VLSI
Implementation of a Wormhole Runtime Reconfigurable Processor,” Master’s Thesis
Virginia Tech June 2001.
[11] JTRS [Online]
Available
[12] SDR Forum [Online]
Available
[13] Hicks, J.E.; Tsai, J.; Reed, J.H.;
Tranter, W.H.; Woerner, B.D., “Overloaded
array processing with MMSE-SIC,”
Vehicular Technology Conference, 2002. IEEE 55th , Volume: 2 , 6-9 May
2002 pp. 542 -546 vol.2.
[14] Zahid, Kazi, Space-time
Processsing for the Wideband-CDMA System, December 2000.
[15] Gozali, R.; Mostafa, R.; Palat,
R.C.; Marikar, S.; Robert, P.M.; Newhall, W.G.; Beaudette, C.; Tsiakkouris, S.A.;
Anderson, C.; Neel, J.; Woerner, B.D.; Reed, J.H.,
“Virginia
Tech Space-Time Advanced Radio (VT-STAR),”
Radio and Wireless Conference, 2001. RAWCON 2001. IEEE , 19-22 Aug. 2001
pp. 227 -231.
[16]
R. Mostafa, K. Dietze, R. C. Pallat, W. L. Stutzman, and J. H. Reed,
Demonstration of real-time wideband transmit diversity at the handset in an
indoor wireless channel,” Vehicular Technology Conference Fall 2001, pp.
2072-2076.
[17] Srivastava, Vikash, Smart
Antennas & Power Management in Wireless Networks,” Master’s Thesis Virginia Tech,
January 2003.
[18] N.J. Drew, and M.M. Dillinger.
“Evolution Toward Reconfigurable User Equipment,” IEEE Communications
Magazine, February 2001, pp. 158 -164.
[19]
P. M. Robert, L. A. DaSilva, J. H. Reed.
Statistcal
Back-Off Method for Minimizing Interference Among Distinct Network Technologies,” Vehicular Technology
Conference Fall 2002.
[20] Neel, James, Jeffrey H. Reed,
Robert P. Gilles. The Role of Game Theory in the Analysis of Software Radio
Networks,” SDR Forum Technical Conference November, 2002.
[21] Neel, James, R. Michael Buehrer,
Jeffrey H. Reed, and Robert P. Gilles. Game Theoretic Analysis of a Network of
Cognitive Radios,” Midwest Symposium on
Circuits and Systems 2002, Volume: 3 ,
Aug. 4-7, 2002 Page(s): 409 -412
[22] Ginde, Samir, Michael Buehrer, and
James Neel, “A Game Theoretic Analysis of the GPRS Adaptive Modulation Schemes,”
Vehicular Technology Conference Spring 2003.
[23] D. Goodman and N. Mandayam. “Power
Control for Wireless Data,” IEEE Personal Communications, April, 2000.
[24] S.H. Wong, and I.J. Wassell,
“Application of Game Theory for Distributed Dynamic Channel Allocation,”
Vehicular Technology Conference Spring 2001.
[25] Jain, Payal. On the Impact of
Channel and Channel Quality Estimation on Adaptive Modulation,” Master’s Thesis
Virginia Tech, December 2002.
There are multiple technological
challenges that need to be overcome to build a software radio. The following
links explain more about SDR related research done at MPRG in the following
areas.
|
