Consensus Distributed Conjugate Gradient Algorithms for Parameter Estimation over Sensor Networks

This paper proposes distributed adaptive algorithmsbased on the conjugate gradient (CG) method and the consensus strategy for parameter estimation over sensor networks. In particular, we present a conventional distributed CG algorithm and a distributed CG algorithm that exploits sparsity in the set of parameters using l1 and log-sum penalty functions. The proposed consensus distributed CG (Consensus- CG) algorithm has an improved performance in terms of mean square deviation (MSD) and convergence as compared with the consensus least-mean square (Consensus-LMS) algorithm and a close performance to the consensus distributed recursive least- squares (Consensus-RLS) algorithm. Similar results are obtained with the proposed sparsity-aware consensus distributed CG algorithm. Numerical results show that the proposed algorithms are reliable and can be applied in several scenarios.