A novel predictive guidance based on generalized predictive control for medium L/D re-entry vehicles

This study introduces a novel predictor-corrector guidance (PCG) law for re-entry vehicles based on generalized predictive control (GPC) to address the convergence and solvability challenges of traditional PCG. The guidance process begins by integrating vehicle dynamics to calculate the terminal range error, which is then embedded into a Controlled Autoregressive Integrated Moving Average (CARIMA) model. This converts the nonlinear equation-solving task in traditional PCG into a system stabilization problem. To address the challenge of significant parameter variations in the CARIMA model, which hinder parameter identification in GPC, a nonlinear transformation method is implemented to narrow the parameter range. This ensures the CARIMA model parameters remain within a smaller, manageable range, enabling accurate guidance command generation in each cycle. The convergence of terminal range errors is rigorously proven using a Lyapunov candidate. A feasibility analysis within the predictive control framework further confirms the method’s ability to resolve solvability issues inherent in traditional PCG. Unlike conventional methods, the proposed approach requires only a single dynamics integration per cycle, significantly improving computational efficiency. Extensive simulations on parachute-assisted landing re-entry vehicles with medium lift-to-drag ratios (L/D) across various flight scenarios demonstrate the algorithm’s accuracy, robustness, and computational efficiency. Comparative simulations verify that versus neural network controllers and iterative PCG, the proposed method achieves superior guidance accuracy with 80% less computational time relative to iterative PCG, making it a promising solution for practical re-entry guidance systems.