Abstract:

In view of the high-dynamic reentry-stage attitude tracking control problem for reusable launch vehicles （RLV） under external disturbances and model parameter uncertainties, an adaptive nonsingular predefined-time convergence control scheme that combines rapid response with simplified parameter tuning is proposed. Firstly, a vary-gain predefined-time convergence form is proposed, integrated with a hyperbolic tangent function to construct a nonsingular vary-gain sliding mode surface and controller, and set the adaptive time-varying gain to optimize controller convergence accuracy. Secondly, a radial basis function neural network （RBFNN） algorithm with low-complexity adaptive laws is developed to compensate for compound disturbances encompassing external perturbations and structural/aerodynamic uncertainties. Finally, through Lyapunov analysis, it can be proved that attitude tracking errors converge to a neighborhood of the origin within a single-parameter predefined time. Explicit relationships are established between the tracking error bounds and parameters of the sliding surface, control law, adaptive neural network, and all signals in the closed-loop system are bounded. Numerical simulations indicate that the method can achieve rapid, high-quality attitude command tracking for RLV.