Optimal Sparse Channel Estimation for Unknown Sparse Level in Wireless OFDM Systems

Abstract

Compressed sensing (CS)has been rapidly adopted as an improved solution to estimate sparse channels for OFDM systems. Nowadays, there exist many algorithms to estimate the channel under this approach. The most popular algorithms belong to the greedy algorithms category. OMP is one of the most useful algorithms due its low computational complexity and good performance. However, the accuracy in this algorithm depends directly on the stopping condition. Currently, there exist many works which try to tackle this problem developing a stopping condition by a threshold. On the other hand, some others works assume this condition known, however in a real scenario it is not true. In this work a simple strategy is proposed permitting to detect the sparse level of the wireless channel and apply it to obtain an optimal solution reducing the MSE in the OFDM channel estimation.