A UAV Trajectory Tracking Method Based on Improved Incremental Model Predictive Control
TL;DR
Simulation comparisons show that AWDC-IMPC achieves lower tracking errors and smoother control than baseline IMPC and the evaluated benchmark controllers, and 200 Monte Carlo trials provide evidence of robustness and performance consistency under bounded uncertainties.
Abstract
Fixed-weight incremental model predictive control (IMPC) has been widely used for quadrotor trajectory tracking because of its predictive capability, but fixed performance weights limit adaptability to trajectory variations, model mismatch, and external disturbances. Existing adaptive MPC methods may require additional iterative parameter optimization or treat weight adaptation and disturbance suppression separately. To address these limitations, this paper proposes an adaptive-weight and disturbance-compensation IMPC (AWDC-IMPC). Unlike existing approaches, AWDC-IMPC coordinates an algebraic multi-factor weight-adjustment rule with residual-based lumped-disturbance compensation in a unified prediction framework. The weights are updated online according to trajectory curvature, position error, and acceleration-correction demand without additional iterative optimization. An LQR attitude controller completes the dual-loop architecture. Simulation comparisons show that AWDC-IMPC achieves lower tracking errors and smoother control than baseline IMPC and the evaluated benchmark controllers. Relative to IMPC, it reduces position-tracking RMSE by 75.72% without disturbances and 86.21% under the tested disturbances. Flight experiments confirm the feasibility and practical tracking performance of the controller, while 200 Monte Carlo trials provide evidence of robustness and performance consistency under bounded uncertainties. These results support the potential of AWDC-IMPC under the tested conditions; however, its performance in more diverse and challenging environments requires further investigation.