Deep Reinforcement Learning-Based Backstepping Control of Air-Breathing Hypersonic Vehicles with Actuator Constraints

This paper investigates the backstepping control problem for air-breathing hypersonic vehicles (AHVs) using deep reinforcement learning (DRL). By utilising the prior knowledge of a widely-accepted AHV model, the proposed DRL-based controller has a high learning efficiency as well as keeps a simple structure. With the help of deep deterministic policy gradient, the intractable parameter adjustment in traditional backstepping designs for AHVs is well handled with full considerations on actuator magnitude and rate constraints. A Monte-Carlo simulation is provided to show the effectiveness of the proposed DRL-based control.