Exponential Stability of LMS-Based Distributed Adaptive Filters

In this work, we consider a class of distributed adaptive filters based on the standard least mean squares (LMS) algorithm, which is proposed to track an unknown signal process in sensor networks. We analyze the stability by introducing a stochastic cooperative information (SCI) condition, in the case of non-independent, non-stationary and possibly unbounded signals. Under the SCI condition, the distributed adaptive filters based on the standard LMS will be shown to be able to track a dynamic process of interest from noisy measurements by a set of sensors working collaboratively, in the natural scenario where any sensor cannot fulfil the estimation task individually. (C) 2017, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.