Iterative decoding/demodulation of coded DPSK systems

A receiver structure suitable for coded M-ary DPSK signals on static and time-varying channels is investigated. The transmitter consists of a convolutional code, an interleaver, and a differential encoder, and hence represents a type of “serial turbo code”. In the receiver, the kernel is an a posteriori probability (APP) demodulator. A priori information is accepted and reliability outputs are produced. The proposed “APP demodulator” uses linear prediction or smoothing techniques and per-survivor processing to estimate the channel response. In combination with a soft-input soft-output decoder, iterative (“turbo”) processing is done, called “turbo-DPSK”. Given a few iterations, on a Rayleigh fading channel the investigated system has the potential to improve the performance of coherent PSK without differential encoding (!) and perfect channel estimation, as opposed to conventional multiple-symbol differential detection techniques. The receiver under investigation can be applied to several existing standards without changing the transmission format