Coupled trajectory optimization and tuning of tracking controllers for parafoil generator

Abstract

ABSTRACT The high-altitude wind power generation (HAWPG) parafoil is an innovative generator with remarkable environmental benefits. Due to the unique trajectory shape, there is an inherent coupling between the pitch and yaw channels, which imposes severe difficulties in the trajectory optimization and tracking controller design. To deal with this intrinsic problem, the trajectory optimization and tracking of the HAWPG parafoil are provided in a comprehensive manner. By using the pseudospectral method (PSM), the parametric optimal trajectory, which has the maximum net power generation, can be obtained. To satisfactorily track this trajectory, well-tuned trajectory controllers are needed to ensure the desired generation performance for the strong coupling dynamics between the two channels. The deep deterministic policy gradient (DDPG) algorithm is applied to the joint tuning of two conventional PID controllers for strongly coupled nonlinear dynamics. A reasonable cost function is established in this process. The mappings between the errors and control gains for the two channels can then be obtained. The gain tuning results are interpretable and well match the empirical knowledge. The simulations are carried out and the comparisons are performed with regard to multiple uncertainties and perturbations. The effectiveness of the proposed method can be illustrated.