Meta-fuzzy observer-based predictive event-triggered fixed-time fault-tolerant control for hypersonic vehicles with compound faults

In this paper, a meta-fuzzy observer-based predictive event-triggered fixed-time fault-tolerant control method is proposed for hypersonic reentry vehicles with time-limited, communication resource-limited, actuator and sensor faults. Considering the compound fault problem, the hypersonic vehicle model is reconstructed into a control-oriented model with sensor and actuator faults. A novel meta-fuzzy fixed-time extended state observer (MFFXESO) is designed. It integrates fuzzy logic with the fixed-time extended state observer to identify the state measurement uncertainty caused by sensor faults and estimate the combined disturbances, including model parameter uncertainties, external perturbations, and faults. Furthermore, a meta-learning framework is employed to optimize the extended state observer gains, enabling rapid adaptation to complex fault scenarios through offline multi-task training and online fine-tuning. To solve the problem of limited communication resources, a predictive dynamic threshold event triggering mechanism (PDTET) is proposed to dynamically adjust the triggering threshold by predicting the future system state, which effectively reduces the consumption of communication resources by the closed-loop controller. Finally, an improved fixed-time non-singular terminal sliding mode surface (NSTSM) is proposed to ensure the tracking error converges quickly in a fixed time to solve the time-constrained problem. The stability and effectiveness of the closed-loop system are demonstrated by Lyapunov analysis and simulation experiments.