Diseño y evaluación experimental de un controlador pid-difuso para el dron dji tello edu

This study proposes a PID-Difuso controller to enhance stabilization of the DJI Tello Edu drone, addressing limitations of conventional PID controllers in dynamic environments. Traditional PID controllers exhibit rigidity when confronted with external disturbances (wind, mass variations) and the inherent nonlinearity of UAVs, resulting in residual oscillations and tracking errors. The proposed solution integrates a two-layer hybrid architecture: • Fuzzy layer: dynamically adjusts PID gains (K_p, K_i, K_d) through heuristic rules based on angular error (e) and its derivative (ė), employing Mamdani inference with triangular membership functions. • PID layer: executes the control law with real-time adaptive parameters. Autonomous flight experiments—including straight trajectories and 180° turns—demonstrated significant improvements over the conventional PID: • RMS yaw error reduced by 23.1%. • MAE yaw error reduced by 25.9%. • Settling time decreased by 28.6%. • Energy consumption decreased by up to 13.1%, extending flight autonomy. • Maximum yaw error during critical maneuvers reduced by 44.4%. The hybrid controller optimizes the trade-off between precision and adaptability. Validation under controlled conditions utilized 20 Hz WiFi telemetry and standardized metrics (RMSE, MAE, control energy). Implementation on a low-cost platform such as the DJI Tello Edu democratizes access to advanced control systems for education and research.