Adaptive control for fault tolerant autonomous launch and recovery

Abstract

This work presents the use of adaptive control methodologies for the recovery to an aircraft carrier of an autonomous fixed wing aircraft. The control architecture used is model reference adaptive control that is implemented in the pitch, roll, yaw, and airspeed axes to provide 6 degree-of-freedom control of the aircraft. Control systems are developed for two aircraft models, a Beech Bonanza representing a low speed high aspect ratio wing aircraft, and the NAVAIR ExJet representing a high speed low aspect ratio wing aircraft. The structure of the controller includes a first-order linear model follower, followed by one of two methods of linear control. The first is a proportional-integral (PI) controller, the second an adaptive critic controller. Adaptation is used to augment the command signal produced by the linear controller, the following methods are used: adaptive bias corrector, optimal control modification, Uadd, and local linear model compensation. An inverse controller, which includes two methods: a force and moment buildup method and a state space model, generates the control effector commands. The control system reference inputs are rotational rates and airspeed, an outer-loop controller is provided to direct the aircraft to the landing point. Control system design is through the use of metrics based on the nominal error, time delay margin, and landing accuracy. The control systems are then evaluated under nominal, effector failure, and modeling error conditions. The control systems defined are capable of providing the desired control under nominal, failure, and error conditions; but will require refinement to improve landing performance.