Neural network‐based fault‐tolerant control for multi‐agent systems with self‐regulation intermediate event‐triggered method

Abstract This paper investigates the consensus tracking control problem for high‐order multi‐agent systems (MASs) subject to actuator faults. A neural network‐based observer and a self‐regulated intermediate event‐triggered (ET) funnel tracking control strategy are proposed to address system uncertainties and enhance fault tolerance. The newly designed observer enables effective estimation of unmeasurable states while reducing the number of neural networks (NNs) required, thereby lowering the computational complexity. Based on this, a fault‐tolerant control (FTC) mechanism is developed to maintain stable consensus tracking performance despite actuator faults. The proposed self‐regulated intermediate ET mechanism effectively eliminates the need for extra compensation functions, minimizes interference with the original control, and enables discontinuous controller updates as well. Additionally, the funnel control technique is applied to prevent excessive control effort. Simulation results on an electromechanical system validate the effectiveness and advantages of the proposed approach.