Neural Adaptive Fixed‐Time Control for Nonlinear Cyber‐Physical Systems With Deferred Asymmetric Time‐Varying Constraints and Malicious Attacks

ABSTRACT This paper proposes a novel neural adaptive fixed‐time tracking control scheme for high‐order nonlinear cyber‐physical systems, which are subjected to deferred asymmetric time‐varying full‐state constraints and malicious attacks on the channel between controller and actuator. A fixed‐time stability criterion for high‐order nonlinear systems is established by innovatively integrating the approximate saturation function with the Barrier Lyapunov Function method. A Gaussian radial basis function neural network is employed to handle the unknown dynamics in the attacked systems while an error shifting function is integrated to resolve the initial value overrun issue through constrained error transformation. The proposed neural adaptive controller achieves enhanced tracking precision, accelerated fixed‐time convergence independent of initial conditions, and rigorous constraint enforcement while effectively mitigating attack impacts, validating its significant control performance improvement over conventional approaches. Rigorous theoretical analysis proves the scheme's capability to guarantee fixed‐time stability and full‐state constraint satisfaction. Practical effectiveness is confirmed through two comparative simulation studies.