Incremental Dual Heuristic Dynamic Programming Based Hybrid Approach for Multi-Channel Control of Unstable Tailless Aircraft

Actor-critic based online reinforcement learning control has been proved to be promising method for of aerial vehicles. However, it is difficult guarantee high-level success rate initial training and tune the large amount parameters actors critics considering unstable multi-input multi-output (MIMO) aircraft. In order facilitate simplify actor critic aircraft, classic stability augmentation system (SAS) designed open-loop aircraft first. Then incremental model dual heuristic dynamic programming (IDHP) method, which proposed recently, extended in application design a multi-channel robust adaptive controller, MIMO form network structures are determined three-channel coupling issues. Consequently, SAS IDHP controller make up novel hybrid framework. this framework, takes charge counteracting eigenvalues system, on guaranteeing performance high-performance tailless equipped with SAS. Specifically, introduction decreases difficulty controller. The tuning process neural networks multiple channels greatly facilitated. Without help SAS, controllers plants almost impossible succeed. Finally, architecture validated using Innovative Control Effectors (ICE) model, modes longitudinal dynamics. Typical aerodynamic uncertainties numerically simulated demonstrate effectiveness method.