Distributed Set-Membership Estimation Over Sensor Networks via an Event-Driven Dynamic Quantization Scheme

Abstract

This article addresses the problem of distributed set-membership estimation for a resource-constrained sensor network. The central aim is to acquire the desired ellipsoidal estimation sets while simultaneously accomplishing improved resource allocation efficiency. Toward this aim, a novel periodic-event-driven dynamic quantization algorithm is developed for each sensor node to save bandwidth on wireless channels and improve measurement accuracy. Such a scheme allows the sensors to implement the quantization process in a dynamic manner. In addition, it conducts a remarkable tradeoff between quantization performance and network energy consumption. Subsequently, a sufficient condition is derived in order to obtain the codesign criterion of the estimator and event-driven scheme using a dedicated auxiliary function. Especially, a recursive convex optimization algorithm is proposed to achieve the suitable ellipsoidal estimation constraint. Finally, the validity of the theoretical results is demonstrated through two illustrative examples.