RL-Based Adaptive Control for Hypersonic Glide Vehicles Under Uncertainty

Hypersonic glide vehicles (HGVs) present significant control challenges due to their highly nonlinear dynamics and the presence of modeling uncertainties and external disturbances. In this paper, we propose a Reinforcement Learning (RL) based adaptive control strategy for HGVs that can achieve accurate trajectory tracking under uncertain conditions. The approach integrates an RL agent with a baseline adaptive controller to compensate for dynamic uncertainties in real time. The RL based controller learns an optimal policy to minimize tracking errors and control effort, while an underlying adaptive mechanism ensures system stability. Simulation results demonstrate that the proposed method can maintain robust performance and stability in the presence of significant aerodynamic parameter variations and disturbances. The approach requires no prior offline system identification, making it suitable for scenarios where accurate models are difficult to obtain.