Online weight design for distributed filtering with limited power

摘要

In this study, the authors consider distributed estimation over a sensor network with limited power. Based on PageRank algorithm, they propose a distributed estimator, where the link weight is designed using online information. They compared the network performance of the proposed estimator with time-varying weight and constant weight under identical initial conditions. To extend the lifetime of sensors, they introduce an online power scheduling method to distributed estimation problem, where each sensor allocates the power for the communication link based on real-time estimation error. They also provide a sufficient condition to guarantee the stability of the networked system with power scheduling.