EM-Based Optimal Maximal Ratio Diversity Combiner for Constant Envelope Signals

Abstract

An optimal maximal ratio combiner (MRC) based on the expectation-maximization (EM) algorithm is devel- oped for noisy constant envelope signals transmitted over a Rayleigh fading channel. Instead of using a transmitted pilot signal with the data to estimate the combiner gains, the EM algorithm is used to perform this estimation. In the developed MRC, estimation of the transmitted data sequence is performed also by the EM algorithm. Estimation using the EM algo- rithm provides an iterative solution to the maximum likelihood (ML) approach. Therefore, the resulting receiver is opti- mum and does not suffer from the difficulties resulted from direct application of the ML procedure. One of these difficul- ties is the computational complexity which depends exponentially on the data sequence length. Introducing an iterative structure in the developed MRC achieves a linear computational complexity and enables efficient data extraction by the Viterbi algorithm when trellis coding is used.