Multi-user Detection Based on Weights Selected Particle Filter in Impulsive Noise

Degradation is a major problem for the particle filter algorithm. Although the conventional resampling method can solve the degradation problem, it will make the particles lose diversity. In this paper, we proposed a weights selected particle filter (WSPF) algorithm for adaptive multi-user detection (MUD) in synchronous code division multiple access (CDMA) system. This algorithm assumes the number of particles required to removal the invalid particles, reserve the particles which make great contributions to state estimation. Thus it can mitigate the degradation of the particle filter. And this algorithm is suitable for the non-linear non-Gaussian system. The simulation results show that the bit error rate of the system can be reduced to a large extent. So the proposed algorithm has important reference value. Key-words: Particle filter, CDMA, Multi-user detection, Weights selected, Non-Gaussian noise