Sparsity-aware sensor selection for correlated noise

The selection of the minimum number of sensors within a network to satisfy a certain estimation performance metric is an interesting problem with a plethora of applications. We have recently explored the sparsity embedded within this problem and have proposed a relaxed sparsity-aware sensor selection (SparSenSe) approach as well as a distributed version of it. In this paper, we generalize our recently proposed sensor selection paradigm to be able to operate even in cases where the measurement noise experienced by the sensors is correlated. We derive the related centralized and distributed algorithms and analyze them in terms of their computational and communication complexities. We also provide general remarks on the convergence of our proposed distributed algorithm. Our simulation results corroborate our claims and illustrate a promising performance for the proposed centralized and distributed algorithms.