Analysis of a DS/CDMA successive interference cancellation scheme using correlations

Compensating for near/far effects is critical for satisfactory performance of DS/CDMA systems. So far, practical systems have used power control to combat the near/far effect. Another approach, still in the research stage, is multiuser detection. In addition to mitigating the near/far effect, multiuser detection has the more fundamental potential of significantly raising capacity by cancelling multiple access (co-channel) interference. There are serious complexity and implementation issues associated with optimal schemes and also with many suboptimal schemes. We analyze a simple interference cancellation (IC) scheme intended to work with power control and related requirements. The IC scheme operates by successively cancelling user interferences ranked in order of received powers. The ranking is obtained from correlations of each user´s chip sequence with the received signal. This is in contrast to an alternative method of ranking based on channel estimates. It is shown that the two methods obtain similar BER performance when channel estimates of the power of each bit have relative errors of around 1 dB if 10 bits are averaged in the correlation method