Distributed Blind Identification Of Sparse Channels In Sensor Networks

Blind channel identification plays an essential role in communications, and various approaches have been proposed in the literature. One of the most important methods for single-input multi-output (SIMO) system identification is the distributed subchannel matching (DSCM) algorithm. However, as the DSCM algorithm treats each component of the channel coefficient vectors equally, it has no advantage when the channels are sparse. In this paper, we propose a kind of sparse DSCM algorithms to blindly identify sparse channels based on the measurements from a sensor network. Unlike the common DSCM algorithm, the proposed algorithm incorporates a sparsity-enforcing regularization term, l(p)-norm (p = 0 or 1) into the cost function to exploit the sparse nature of channels. Several simulations are then presented to show that the proposed algorithm can improve the performance of channel identification in both convergence and accuracy.