Iterative Receiver Design With Off-the-Grid Sparse Channel Estimation

Recent progress in wireless receiver design has been towards iterative processing, where channel estimation and decoding is considered a joint optimization problem. Sparse channel estimation is another recent advancement, which exploits the inherent structure of those wireless channels that are composed of a small number of multipath components. In this work we design iterative receivers which incorporate sparse channel estimation. State-of-the-art sparse channel estimators simplify the estimation problem to be a finite basis selection problem by restricting the multipath delays to the discrete domain (i.e. to a grid). Our main contribution is a receiver without such a restriction; the delays are estimated directly as continuous values. As a result, our receiver does not suffer from the leakage effect which destroys sparsity when the delays are restricted to the discrete domain. We discuss certain connections between continuous and discrete domain sparse estimation methods. Our receivers outperform state-of-the-art sparse channel estimation iterative receivers in terms of bit error rate.