Consensus-Based Robust Filtering Over Sensor Networks With Weighted and Unweighted Topologies and Communication Delays

In this article, a consensus-based robust filter is proposed for sensor networks with weighted and unweighted topologies, and communication delays. The system is considered to be linear time-varying (LTV) with bounded uncertainties. In order to achieve the desired filter, first, a robust optimization problem is defined based on the consensus protocol for unweighted topologies. The consensus protocol ensures an agreement between the estimates of all sensors. By solving this problem, the proposed consensus-based robust filter is obtained and its algorithm is presented. Then, to increase the applicability of the filter, another optimization problem for weighted topologies is introduced. Next, due to the impact and high importance of communication delays in sensor networks, by considering the time-varying delays in consensus protocols, new optimization problems are solved again to achieve a suitable filter. One of the most significant features of the proposed consensus-based robust filters is that the loss of a large part of the network links does not pull down the filtering performance. Finally, in the simulation section, the robust performance and efficiency of the proposed consensus-based robust filters have been investigated for weighted and unweighted networks in the presence of time-varying delays and loss of communication links.