Joint Channel Identification and Estimation in Wireless Network: Sparsity and Optimization

In this paper, we study channel identification for a wireless network with the aid of compressed sensing (CS) in both cases of known and unknown sparsity levels of clusters. For the unknown case, we propose using blind CS signal recovery algorithm to sequentially estimate both sparsity level and channel gains. The refined version of blind CS technique is also provided to improve the consistency of channel identification process. The convergence of two algorithms is guaranteed by ensuring that the cost functions decrease after each update. We then investigate the cluster sparsity of users in the case of known sparsity levels of all clusters. By exploiting the alternating direction method of multipliers algorithm through distributed optimization, we can identify channels in sparse clusters parallelly and efficiently compared with conventional convex techniques. In summary, this paper provides some insight into employing sparse and distributed algorithms to efficiently solve the problem of fast channel identification and estimation of users in future network.