Performance of a Novel Adaptive Multistage Full Parallel Interference Canceller for CDMA Systems

In this paper, we investigate the statistical behavior of the adaptation partial cancellation weight (PCW) in partial parallel interference cancellation (PPIC) schemes based on least mean square (LMS) algorithms in which the adaptation PCW of the last chip in a symbol period reveals the information of the correctness of the bit decision for direct-sequence code-division multiple-access (DS-CDMA) communication systems. We obtained an adaptive decision threshold using a maximum a posteriori probability (MAP) detection criterion for the PCWs to judge the correctness of the bit decision. A bit-inversion procedure is proposed to invert the incorrect bit decisions. After bit inversion, full PIC is performed to cancel multiple-access interference (MAI). Simulation results show that the proposed three-stage LMS-based adaptive full PIC (AFPIC) multiuser detector reaches the performance of a single user bound over flat fading channels. Moreover, the AFPIC detector effectively cancels MAI over a two-path frequency-selective fading channel.