Distributed Fault Detection for Formation of Multi-agent Systems

In this paper, the distributed fault detection (FD) problem is addressed for a class of linear stochastic multi-agent systems (MASs). A novel distributed cooperative control scheme is given for reaching bounded formation of the MAS. It is proved that by using relative outputs between neighboring agents, a set of distributed fault detection filters can be designed for each agent to detect the faults occurring in its general neighbor set. The recursive design protocols of the filter parameters are presented to guarantee minimum variance estimation. The residual generated by each filter is robust against one neighbor's fault while sensitive to that of other neighbors. Finally, a simulation study on a practical unmanned aerial vehicle (UAV) formation is provided to demonstrate the effectiveness of the approach proposed in this paper.