Iterative decoding for coded noncoherent MPSK communications over phase-noisy AWGN channel

A family of interleaved systems for communicating over the additive white Gaussian noise (AWGN) channel with noisy or unknown phase is investigated. Each system comprises convolutional code or turbo code, interleaver, binary phase shift keying or quaternary phase shift keying modulation and differential or similar encoding. The receiver performs jointly and iteratively the decoding and the channel demodulation in the presence of unknown phase utilising an observation interval exceeding two symbols. All systems exhibit good performance with various levels of phase noise. The convolutional code, achieving bit error rate (BER) below 10-3 at bit energy over noise spectral density ratio (Eb/No) of 2.4 dB with no phase noise, outperforms the turbo code slightly over AWGN channels due to limited capability of the iterative algorithm in this setting. The turbo code is superior in bursty channels, modelling for example a block erasures fading environment. When the carrier phase is unknown but constant over an interval of ten symbols, the turbo coded system achieves BER lower than 10-3 at Eb/No of 2.6 dB, that is 1.3 dB away from the capacity limit of this channel.