High velocity path control of quadrotors

This paper presents a new structure for path following control that allows multirotor UAVs to fly on predefined paths with high velocity. The goal is to maximize the control bandwidth of the quadrotor, yet with good robustness against actuator saturation, model uncertainties and aerodynamic disturbances at high velocity. The suggested control approach is composed of a path-based reference model with in-flight adjustable path velocity and a Nonlinear Dynamic Inversion (NDI) trajectory controller with L₁ adaptive augmentation. In previous control designs, only up to the second derivative of the path reference (i.e. reference acceleration) is used in the feedforward control. The suggested NDI control structure enables feedforward control of the third order derivative of the reference trajectory (i.e. reference angular rate of quadrotor), so that the control bandwidth can be improved. Yaw rotation remains as an additional degree of freedom. Experimental flight tests are performed on a quadrotor with GPS receiver and low cost inertial sensors for state estimation. They demonstrate high velocity path following in the presence of thrust saturation, external disturbances and modeling uncertainties.