Parity aided detection of SM-MIMO systems

Abstract

For spatially multiplexed MIMO systems, conventional detection methods such as MLD and QRD-M requires high computational complexity when the number of antennas or modulation level increases. This paper presents a method for reducing the computational complexity with the aid of parity information sent through one of the transmitted symbols. The method utilizes QR decomposition of the channel matrix as in conventional QRD-M, but, instead of retaining M candidate symbols at each level of tree decomposition, only one candidate is kept, which reduces the computational complexity. The probability of error in the selection of the candidate to be retained at each level is kept to a minimum with the aid of parity information. Simulation results show that compared to QRD-M of same performance, the computational complexity is reduced by 74%. In terms of BER, an improvement of 3dB is achieved even at low SNR levels. Compared to coded systems also, an improvement in BER is achieved, with much less decoding complexity.