Distributed incremental-based rls for node-specific parameter estimation over adaptive networks

In this paper, we introduce an adaptive distributed technique that attains the exact (Recursive Least Squares) RLS solution of a node-specific parameter estimation problem where each node is interested in a set of parameters of local interest and a set of global parameters. To do so, each node of the network relies on its own local data as well as the communication with its intermediate neighbor under an incremental mode of cooperation. Since the required inter-node communication of the new scheme may be prohibitive for networks with scarce energy resources, an alternative low-cost scheme is derived to reduce the communication burden. It is shown that this approximate strategy may attain the exact RLS solution in the steady state. To illustrate the effectiveness of the proposed techniques we provide some indicative simulation results.