Trajectory Tracking Controller Design Using Neural Networks for Tiltrotor UAV

Abstract

This paper presents a design of trajectory tracking controller for Smart UAV, a tiltrotor unmanned aerial vehicle . Its design objective is to control the tiltrotor aircraft with one fixed controller architecture for all flight modes, that is, helicopter mode , conversion mode, and airplane mode . The controller is implemented u sing two time -scale separation and thus composed of inner -loop for attitude control and outer -loop for trajectory control. The dynamic model inversion technique is applied to both control loops such that the inner -loop DMI is driven using the angular equat ion of motion and the outer -loop DMI using the translational equation of motion. In additi on on -line adaptive neural network s are employed to compensate the model inversion error s of the inner -loop and the outer -loop DMIs, due to the lack of full knowledg e of tiltrotor aircraft dynamics. And the pseudo -control hedging al gorithm is adopted to lessen instability problem caused by the time -delay of control loops and saturation and rate limit of actuators. The trajectory tracking controller is evaluated usin g a sophisticated 6DOF nonlinear simulation program with an approach and landing scenario including all flight modes of the Smart UAV .