Fault estimation for networks of non-homogeneous agents with switching topologies

摘要

In this study, the problem of distributed fault estimation for multi-agent systems with switching topolo-gies is investigated. Firstly, a "switched virtual model" based on local information is used to represent the model of each agent. Then, some coordinate transformations are proposed to decouple a sub-state of the virtual model from faults. Next, a local impulsive switched observer is designed for each agent in order to estimate the state. Sufficient conditions for the solvability of the observer design are provided in LMIs form. Finally, the faults are estimated by combining the estimated states and a robust exact differentiator based on sliding mode technique. In this proposal, each agent estimates its own faults and those of its neighbors. The effectiveness is illustrated by a simulation example of a team of Unmanned Aerial Vehicles (UAVs). (c) 2020 European Control Association. Published by Elsevier Ltd. All rights reserved.