Energy-Efficient Trajectory Optimization with Nonlinear Model Predictive Control for a Quadrotor UAV

This paper proposes a control law that improves the energy efficiency of a quadrotor UAV. An optimal velocity hypothesis was followed, thanks to the correlation between ground velocity and total consumed energy. An explicit energy model is developed and employed within the trajectory optimization problems. The proposed approach comprises an optimization stage at which an energy-efficient trajectory is generated followed by a Nonlinear Model Predictive Controller that tracks the generated optimal position and velocity subtrajectories, while the open-loop optimal control problem used to generate the trajectory is dynamically constrained. The proposed control approach was extensively tested in simulation and compared to tracking obtained with standard NMPC, PID, and LQR feedback controllers. The preliminary results demonstrate the validation of our hypothesis and a noteworthy reduction in energy consumption when compared with the use of the other standard feedback controllers, a 36% of energy saving was achieved in some cases.