Transmission of Non-Uniform Memoryless Sources over Wireless Fading Channels via Non-Systematic Turbo Codes1

In this work, we investigate the joint source-channel coding problem of transmitting non-uniform memoryless sources over wireless Rayleigh fading channels via Turbo codes. The source redundancy in the form of non-uniformity is exploited in the Turbo decoder by incorporating the source statistics in the modified extrinsic information. In contrast to previous work, non-systematic recursive convolutional encoders are proposed as the constituent encoders, as they produce almost uniform outputs irrespective of the degree of non-uniformity in the source. As a result, unlike the outputs of systematic encoders, they are suitably matched to the channel input since a uniformly distributed input maximizes the channel mutual information and hence achieves capacity. Simulation results show substantial gains achieved over previously designed systematic Turbo codes, and the gaps to the optimal Shannon limit are therefore significantly reduced. In comparison with a tandem scheme, our system offers robust and superior performance at low BER levels (below 10 4), while the complexity is lower.