Performance of iterative data detection and channel estimation for single-antenna and multiple-antennas wireless communications

In this paper we present analytical results on the performance of an iterative data-detection/channel-estimation scheme, with reference to both the situations of a single-antenna and of a multi-antenna communication system. The results of the analysis include the derivation of closed-form formulas for the channel estimation mean square error at any given iteration, and for its Cramer-Rao bound, as functions of the error probability. Moreover, the problem of the optimal choice of the length of the training sequence is also addressed, and a graphical proof of the convergence of the proposed algorithm is also presented. Extensive computer simulation results confirm the huge performance gains granted by the iterative strategy and demonstrate the accuracy of the analytical formulas.