Maximum-likelihood differentially coherent communication with multiple-symbol observation interval

Abstract

In this dissertation work, several differentially coherent detection algorithms for multiple phase shift keying(MPSK) signaling are proposed for enhancing the performance of the classical differential detection(DD). The proposed algorithms can recover a portion of the performance loss resulting from DD compared to that of coherent detection(CD) and yet, still maintain a simple but robust implementation. CD is optimum demodulation for MPSK signaling over the additive white Gaussian noise(AWGN) channel. In many cases, e.g., in a fading environment where carrier recovery is difficult, differentially coherent detection of differentially encoded signals is preferred [1],[2]. Although differential detection eliminates the need for carrier acquisition and tracking, it suffers from performance degradation compared to ideal CD. The amount of this performance penalty increases with the number of phases M and is significant for M ≥ 4[2]. D. Divsalar and M. K. Simon[3] suggested the multiple-symbol differential detection(MSDD) algorithm, which has better performance than that of the classical DD with some additional complexity. The MSDD algorithm makes a joint decision on several symbols simultaneously which is a form of maximum-likelihood sequence estimation. In this work, several novel differentially coherent detection techniques, which use multiple-symbol observation interval, are proposed and their performances are analyzed and obtained by simulation. With these schemes, a complicated phase-tracking circuitry and the associated phase acquisition time are not needed. In the proposed maximum-likelihood multiple differential detection algorithm, the differential vector is formed by rotating the received sequence based on the previously received signal to compensate the arbitrary phase introduced in the channel. This signal vector is accumulated with a phase adjustment to minimize the distance between this vector and the accumulated vector. This is the maximum-likelihoo...