Sparse system identification over adaptive networks

Recently several algorithms have been developed to exploit the distributive nature of an ad hoc wireless sensor network to estimate a certain parameter of interest. However, none of the proposed algorithms addresses the issue of sparse system identification over an adaptive network. Recently, several LMS-based sparse estimation algorithms have been devised. This work proposes a distributed sparse LMS algorithm for parameter estimation over adaptive networks. Two different distributed schemes have been used for incorporating the sparse LMS algorithm into the adaptive network framework. Mean transient analysis has been carried out. Simulation results show that in a sparse environment the proposed algorithms perform better than the LMS algorithm.