A linear minimum mean square error multiuser receiver in Rayleigh fading channels

In this paper, we generalized the multiuser communications system to fading environments. We first derive the Verdu's conventional optimum receiver and then evaluate its performance. To efficiently estimate the received signal strength of each user, a linear minimum mean square unbiased estimator (LMMSE) is used. Applying the minimum mean square error (MMSE) Bayesian estimation, we further propose the LMMSE bit estimator which can efficiently demodulate the information bits. We demonstrate its close performance to the optimum multiuser receiver but with much simpler polynomial complexity. Furthermore, we extend our LMMSE estimator to the sequential LMMSE estimator. In sequential estimation, we do not need to implement the matched filters and perform the matrix inversion when estimating. In addition, the convergence speed is approximately 2k times of iterations; where k is the number of users. Moreover, the sequential LMMSE estimator can be easily generalized to the asynchronous transmission case.