A full authority helicopter adaptive neuro-controller

This paper treats the development of a full authority, six degree of freedom controller for rotorcraft which provides autonomous, high performance, robust tracking of a specified trajectory. The nominal controller is a two time scale input-output-linearizing controller which exploits the well known nonlinearities in the equations of motion, but ignores the variations in the aerodynamically varying quantities. The nominal controller is enhanced with a simple two layer adaptive neural network which accommodates for the variations in the dynamics and guarantees ultimate boundedness of the tracking errors in closed loop. Simulation results are presented employing a high fidelity simulation for the Apache helicopter which has been validated at several flight conditions against flight test data.