Data-Driven Tracking Control for Nonaffine Yaw Channel of Helicopter via Off-Policy Reinforcement Learning

This article presents an off-policy tracking control scheme for the continuous-time nonaffine yaw channel of uncrewed aerial vehicle helicopter. First, the article constructs an affine augmented system (AAS) within a parallel control structure to convert the original nonaffine tracking error dynamics into affine dynamics. Second, the article derives a stability criterion linking the nonaffine system and the AAS, demonstrating that the obtained zero-sum policy from the AAS can achieve the <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$H_\infty$</tex-math></inline-formula> performance of the nonaffine system. Third, a data-driven off-policy tracking algorithm is designed for approximating the zero-sum solution of the Hamilton–Jacobi–Isaacs equations with unknown dynamics. Moreover, the recursive least squares process with a variable forgetting factor is employed to update the actor-critic neural network weights, with the algorithm's convergence being proven. Then, the uniformly ultimately bounded of tracking errors is guaranteed. Finally, two application examples are offered in simulation to validate the effectiveness of this presented method.