Abstract:

To solve the problem of symbol detection in the presence of the unknown carrier phase, a Bayesian detection algorithm is proposed based on the Gibbs sampler method. The proposed algorithm draws random samples from the conditional posterior distribution of all unknown quantities, so that the joint symbol detection and carrier phase recovery is accomplished. A salient feature of the detection algorithm is that it has a softinput softoutput structure. Therefore, it is well suited for iterative processing in a coded communication system, which allows the Bayesian symbol detector to improve its performance. Simulation results show that the proposed algorithm outpeforms the traditional nondataaided (NDA) algorithm significantly. In a Turbo code system, the bit error rate performance gain of the proposed algorithm is more than 1 dB compared with the NDA algorithm.