Cooperative Fault Estimation for A Class of Heterogeneous Multi-agents with Stochastic Nonlinearities Based on Finite Impulse Response Filter

This paper investigates the cooperative fault estimation problem for a class of heterogeneous multi-agent systems, in which the agent dynamics are governed by linear discrete time-varying systems with nonidentical dimensions subject to stochastic nonlinearities. A finite impulse response (FIR) filter based fault estimation scheme is developed via relative outputs to estimate the possible faults of the local and neighboring agents simultaneously. An analytical redundancy expressed in terms of all the states in the previous time window is originally established for deriving the fault estimation signal. The prior variance information coupled with fault estimation error in nonlinear form, is fully considered to design performance index through analysis of random matrix inequality. The optimal FIR filter gain is analytically obtained with computational efficiency by searching the minimum point of the relevant matrix trace function. Illustrative examples are finally provided to demonstrate the effectiveness and advantages of the developed results.