Detection method for MC-CDMA based on a recurrent neural network structure

A detection method based on a recurrent neural network structure is derived for a multi carrier code division multiple access communication system with multi path propagation. Contrary to other neural network approaches, the RNN has the advantage, that network size as well as the coefficients of the network can be derived from parameters which characterize the communication system. The energy function of the RNN is identical to the log-likelihood function of the maximum likelihood detector. Different iteration algorithms for the RNN with an emphasis on parallel processing are discussed. Performance results are given for the Rayleigh fading channel and a typical mobile radio channel. Performance and complexity of the RNN detector are compared with other iterative detection algorithms, specifically a block decision feedback equalizer.