Implementation of an SDR platform using GPU and its application to a 2 × 2 MIMO WiMAX system

Conventional communication systems have been implemented using digital signal processors (DSPs) and/or field programmable gate arrays (FPGAs), especially for software defined radio (SDR) functionality. We propose a scheme that uses a graphics processing unit (GPU) in place of the conventional DSPs or FPGAs for the implementation of an SDR-based communication system. The GPU, a high-speed parallel processor with multiple arithmetic logic units, is adopted for the signal processing of the physical layer required for the parallel processing in an SDR system. The compute unified device architecture (CUDA) based on the C language provides a software development kit (SDK) for the modem application of the GPU. Therefore we utilize the CUDA SDK to implement the real-time modem function. This paper presents an implementation of a 2 × 2 multiple-input multiple-output (MIMO) WiMAX system employing a GPU as the real-time modem. By installing a radio frequency module on top of the GPU modem, we implement a real-time transmission system for video data. The performance of the proposed GPU-based system is demonstrated by comparing its operation time against that of the conventional DSP-based system.

[1]  Johan Eilert,et al.  VLSI implementation of A multi-standard MIMO symbol detector for 3GPP LTE and WiMAX , 2010, 2010 Wireless Telecommunications Symposium (WTS).

[2]  Seungwon Choi,et al.  Implementation of an SDR system using graphics processing unit , 2010, IEEE Communications Magazine.

[3]  Ramesh Annavajjala,et al.  BER analysis of QAM on fading channels with transmit diversity , 2006, IEEE Transactions on Wireless Communications.

[4]  John D. Owens,et al.  GPU Computing , 2008, Proceedings of the IEEE.

[5]  Erik Lindholm,et al.  NVIDIA Tesla: A Unified Graphics and Computing Architecture , 2008, IEEE Micro.

[6]  Per Ola Börjesson,et al.  ML estimation of time and frequency offset in OFDM systems , 1997, IEEE Trans. Signal Process..

[7]  Sinem Coleri Ergen,et al.  Channel estimation techniques based on pilot arrangement in OFDM systems , 2002, IEEE Trans. Broadcast..