Finite-time fault-tolerant cooperative control for multi-agent systems with input saturation and unknown control coefficients

This paper addresses the cooperative tracking problem for high-order uncertain nonlinear multi-agent systems under complex conditions including non-affine faults, input saturation, unknown control coefficients, and external disturbances. A novel finite-time adaptive fault-tolerant control strategy is proposed based on the command-filtered backstepping control framework. Specifically, RBF neural networks are employed to effectively approximate and suppress the effects of unknown nonlinear dynamics caused by non-affine faults and external disturbances, while an improved adaptive mechanism is developed to significantly reduce computational complexity. To resolve control input saturation and unknown control coefficients, a Nussbaum-type function and a novel gradient regulator design are incorporated into the controller architecture, ensuring control efficacy under saturation constraints while avoiding numerical instability. Furthermore, based on the finite-time control technique, the closed-loop system signals are guaranteed to rapidly converge and remain bounded within finite time. Finally, two comparative simulation experiments validate the controller’s performance, demonstrating that the proposed algorithm achieves satisfactory tracking even in fault scenarios.