Maximizing Network Lifetime for Event-Triggered Distributed Estimation with Performance Constraint

We consider the problem of maximizing the lifetime of a wireless sensor network (WSN) deployed for distributed estimation of a time-varying parameter subject to a meansquared deviation constraint. By employing a set-membership filtering (SMF) approach that features event-triggered updates of parameter estimates, we reduce the energy consumption in local processing and transmission through sparse updates of the parameter estimates. In this SMF event-triggered framework, the objective of maximizing the network lifetime (NLT) can be naturally tied to the estimation performance of the network. Accordingly, the NLT maximization problem is posed as one of non-convex optimization problems. We propose an iterative algorithm based on binary search that yields a globally optimal solution. The optimization results in significantly increased NLT compared to conventional filtering methods such as normalized least mean-squares (NLMS) algorithm.