Sequential reconstruction of vector quantized signals transmitted over Rayleigh fading channels

The paper presents an optimal sequential decoding scheme for a joint source and channel coding system operating in the Rayleigh flat fading channel. The minimum mean-square error (MSE) between the original and the reconstructed source signals is used as the optimality criterion. The system being investigated consists of a vector quantizer (VQ) whose output indices are mapped directly into points in the modulation signal space. The modulation signal is then transmitted over a Rayleigh fading channel. A sequential decoder based on the Bayesian estimate is used to reconstruct the source signal from the received signal samples and from the channel state information (CSI). A recursive algorithm for implementing the Bayes receiver is introduced. Compared to the symbol by symbol decoding technique proposed earlier, it is found that sequential decoding significantly improves the system performance when a correlated source is transmitted.<<ETX>>