Analysis of adaptive multistage interference cancellation for CDMA using an improved Gaussian approximation

We consider a simple model for adaptive multistage interference cancellation within a CDMA system, and seek to develop an accurate analytical expression for the performance of this system. Previous work on interference cancellation has relied heavily on simulation techniques or a simple Gaussian approximation (GA). The standard GA ran lead to bit-error rate (BER) results which are optimistic for the conventional receiver, and this also occurs when the approximation is applied to the interference cancellation problem. Additionally, this approximation does not allow the second order effects of the multiple access interference (MAI) to be included in the performance estimates. Several improvements on the standard GA have been suggested which result in accurate performance results for a standard CDMA receiver. This paper presents an analytical expression for the probability of bit error for an adaptive multistage interference canceller, using an improved Gaussian approximation (IGA) for MAI. The BER at any stage of interference cancellation can be recursively computed from the signal-to-noise ratio (SNR), the statistics of the random powers of users, and the processing gain of the CDMA system. The performance of the resulting EER expression is compared with simulation results. Since the second order effects of MAI can be included, the analytical framework presented here can also be used to evaluate the performance of multistage interference cancellation in arbitrary fading environments, and we present results for the performance of interference cancellation in lognormal fading environments