UAV performance motion control and wind disturbance rejection based on PPO algorithm and controller integration

Accurate control of drone movements is crucial during the performance process. However, existing control methods have problems such as insufficient control accuracy and limited ability to resist wind and interference. This study aims to address these issues by enhancing trajectory tracking precision and wind-disturbance rejection capabilities. Consequently, this study proposes an integrated flight control method that combines the Proximal Policy Optimization algorithm with a controller network. Within this framework, the Proximal Policy Optimization network dynamically optimizes the control signals of a proportional-Integral-Derivative controller through environmental interactive learning, thereby improving tracking accuracy under calm conditions. In the presence of wind interference, the real-time calculation of wind disturbance compensation is added to the controller output to achieve precise control of environment adaptation. Simulation results demonstrate that under zero-wind conditions, the proposed method yields a maximum positioning error of merely 0.015 m, while maintaining tight alignment with the desired trajectory in both straight and curved segments (maximum error: 0.021 m). Subject to a 5 m/s constant wind, the proposed framework minimizes positioning errors across all five target waypoints, bounding the maximum deviation within 0.099 m. Under abrupt wind gusts, the maximum positioning error reaches only 0.159 m, which is significantly lower than those of the three benchmark methods. This study indicates that the proposed method exhibits clear advantages in trajectory precision and robustness, providing robust technical support for highly reliable Unmanned Aerial Vehicle formation flights.