EM-based channel estimation for cooperative relay networks

We consider channel estimation in amplify-and forward (AF) wireless relay networks (WRNs) when coherent distributed space-time coding is performed at relays. We perform minimum mean-square-error detection at the destination when partial channel state information is available. For channel estimation, we propose to use a semi-blind method based on the expectation maximization (EM) algorithm. Using an iterative channel estimation scheme, we show that a considerable improvement can be achieved in signal detection after a few iterations. We discuss the optimality of the proposed estimation scheme and further propose to improve its performance by combining appropriately the channel estimates based on pilot and data symbols.