Adaptive Fixed-Time Active Fault-Tolerant Control for Nonlinear Systems Based on Super-Twist Sliding Mode and RBFNN

This paper proposes a novel fixed-time active fault-tolerant control method for second-order affine nonlinear systems. The key innovation lies in the integration of a super-twisting barrier function based adaptive sliding mode control (ST-BFASMC) with a radial basis function neural network (RBFNN) observer, achieving simultaneous improvements in convergence speed and fault tolerance. Firstly, a RBFNN observer with weight and centre value update strategy is designed. Secondly, a non-singular fast terminal sliding surface and control law with a super-twisting term are constructed. Furthermore, the fixed-time convergence properties of both the controller and observer are rigorously proven using Lyapunov stability theory. Experimental studies on quadrotor UAV attitude control demonstrated that the adopted RBFNN observer achieved over 60% improvement across all performance metrics compared to baseline methods, while the control algorithm exhibited more than 30% enhancement in multiple indicators relative to conventional ASMC and barrier function based ASMC (BFASMC) approaches. These results validate the algorithm's strong robustness and fault-tolerant capability in the presence of actuator failures.