Fixed-time adaptive neural network fault-tolerant control during the Belly Flip maneuver for a starship-like vehicle

A fixed-time nonsingular terminal sliding mode fault-tolerant controller (FxNTSMC-FTC) is developed for a starship-like vehicle’s ”Belly Flip” maneuver. Different from the previous work, a holistic control scheme that couples aerodynamics, trajectory, and control is presented, achieving rapid and precise large-angle attitude stabilization under multiple uncertainties and actuator faults. First, the actuator model and trajectory optimization model are proposed to calculate the aerodynamic parameters and the reference trajectory, and the attitude dynamics model is established using quaternions to prevent the singularity phenomenon. Then, a fixed-time adaptive radial basis function neural network (RBFNN) disturbance observer (RBFNN-FxTDO) are introduced to handles uncertainties and disturbance, which effectively estimates internal uncertainties caused by mass and inertia changes and external disturbance such as aerodynamic deviation and unmodeled deviation. Finally, a fixed-time nonsingular terminal sliding mode fault-tolerant controller is developed to address the large-angle attitude maneuvers during the short-duration ”Belly Flip” process, ensuring rapid, precise, and robust attitude tracking, with rigorous fault tolerance and fixed-time stability. Simulation results demonstrate the scheme’s effectiveness and superiority.