A Modified Sphere-Decoding Based Detector for MIMO LS-CDMA System

Loosely-synchronized code division multiple access (LS-CDMA) technology is an excellent CDMA technology and very suitable to be applied in the future wireless communication systems. When the multiple input multiple output (MIMO) technique is integrated, better performance and system capacity can be achieved. But the computational complexity of classical detection algorithm based on exhaustive searching, such as maximum-likelihood (ML) rule, is too high and inconvenient to realisation. Although sphere-decoding (SD) algorithm is able to greatly decrease the detection complexity due to the sphere-constrained searching, the decoding complexity is very high especially when the constellation has a lot of points and the number of antennas is large. A modified detection algorithm based on SD method is proposed in this paper, for the sake of further decreasing the complexity. It decodes according to the signal-noise-ratio (SNR). The detector outputs the zero-forcing (ZF) detection results when the SNR is relatively high, otherwise it subtracts these decoded signals from the original receiving signals and then decodes them by SD algorithm. The simulation results shows that the performance of such modified algorithm has little difference as compared with that of SD algorithm.