Disturbance Estimation and Compensation Based on Inverse Model Learning for Quadrotor Robust Control
摘要
This paper proposes a Back Propagation (BP) neural network-based feedforward-feedback control method to improve quadrotor anti-disturbance performance. The approach uses BP networks to approximate the inverse model of quadrotor system. So we can use this BP network and known system inputs to estimate external disturbances without requiring exact system parameters, enhancing adaptability and robustness. Simulations show the BP network outperforms Radial Basis Function (RBF) Network in disturbance estimation and verify the control performance advantage of the proposed method.