Improved iterative parallel interference cancellation for coded CDMA systems

A novel method for interference estimation in the iterative parallel interference canceller is presented. The standard log-likelihood ratio algorithm suffers from a bias problem that limits the system capacity. We propose an improved iterative parallel interference canceller which is able to overcome this problem. In a highly loaded system the correlation between the the decoder inputs in successive iterations is low enough so that a simple combining of the decoder inputs, as if they were independent, yields a high capacity increase.