Iterative noncoherent detection of differentially encoded M-PSK

The combination of convolutional encoding, interleaving and differentially encoded M-ary phase shift keying (M-PSK) forms an attractive transmission scheme because it combines the benefits of error correction with the simplicity of noncoherent detection. The use of an interleaver, which rearranges the bits of the code word prior to modulation, has the potential to greatly enhance the error correcting ability of the code, provided a suitable decoder structure is used. In this thesis an iterative decoder structure, similar to the one used for decoding "turbo" codes, is proposed for use with this transmission scheme. The resulting technique, named iterative noncoherent detection, is applied to three different, well-known channel models; namely, the additive white Gaussian noise channel, the Rayleigh frequency-flat fading channel, and the Rayleigh frequency-selective fading channel. For each channel model novel soft-output decoders are derived, and the system performance, in terms of the bit error rate, is investigated by means of computer simulation. In all three applications, the proposed system works very well, not only providing performance that is significantly better than traditional non-iterative techniques, but also much better than anticipated.