Optimum receiver design for correlated Rician fading MIMO channels with pilot-aided detection

Two receiver structures based on pilot symbol-aided channel estimation are considered for the separately-correlated Rician fading MIMO channel. 1) A mismatched receiver, which decodes the received signal by first using maximum-likelihood (ML) or minimum mean-square error (MMSE) estimation of the MIMO channel matrix, and then by assuming that the estimate is exact. 2) An optimum receiver, which does not estimate explicitly the channel matrix but jointly processes the received pilot and data samples assuming known channel distribution. The main focus is on the optimum receiver. First, the optimum detection algorithm for the separately-correlated Rician fading MIMO channel is derived. Then, an iterative implementation suitable for trellis space-time decoding is proposed in order to reduce the algorithm complexity. Numerical results are presented for a 2 x 2 MIMO system with Rayleigh/Rice correlated/uncorrelated fading and a simple trellis space-time code. These results show that substantial gain is available by using our proposed optimum receiver either in terms of Eb/No and system throughput (in both cases accounting for pilot-symbol rate reduction). Finally, the effect of parameter estimation (required by the optimum receiver) is studied and it is shown that this effect is almost unnoticeable in the case considered.