Joint Symbol Detection for Multi-Receiver Without Signal Synchronization and Array Alignment

Abstract

This letter deals with symbol sequence detection of multi-receiver signals based on the maximum-likelihood (ML) criterion. A novel iterative scheme is proposed to estimate the symbols without synchronization or alignment of timing and phase. Instead of the accurate delay compensation and optimal samples recovery, the detection of symbols employs the direct calculation of the matched filter outputs, eliminating the need for separate synchronizer and combiner. Since the received samples at each branch are viewed as incomplete observations, the joint ML estimation of symbols can be obtained in an iterative way by means of the expectation-maximization algorithm. The proposed scheme is compared with conventional signal combining schemes and joint ML synchronizers by simulation results. It is shown that the proposed detector enables significant improvement on the bit-error rate for weak received signals.