Model Predictive Control for Rotor Failure of a Quadcopter Based on Disturbance Observation

This study provides a Disturbance-Observer-based Model Predictive Control (DO-MPC) method to address the stability issues of quadcopters under rotor failure. First, a nonlinear dynamic model of the quadcopters is established, and a DO method is introduced to estimate environmental disturbances and actuator faults in real time. The estimated disturbances are then incorporated into the MPC optimization process as compensation terms. To reduce computational complexity, a stepwise control optimization strategy is further proposed. Finally, experiments comparing DO-MPC with the Lyapunov method demonstrate that DO-MPC maintains flight stability under single rotor failure while achieving faster response, smaller control variations, reduced attitude oscillations, and better suppression of environmental disturbances.