Pre-processed back propagation neural networks for CDMA interference cancellation

Abstract

This paper attempts to show the effects of pre-processing in bit error rate of neural network interference cancellation in spread spectrum receivers. Multiple access interference (MAI) which is one of the primary issues in CDMA multiuser detection is combated by interference cancellers. Feed-Forward Neural Networks with back propagation is used for interference cancellers. The works shows that pre-processing can bring adequate improvement in bit error rate as well as complexity.