Dual-Authority Thrust-Vectoring of a Tri-TiltRotor employing Model Predictive Control

This paper addresses the exploitation of the combined potential of the directly-actuated and the underactuated control authorities of unmanned aerial vehicles with thrust-vectoring actuation. For the modeling, control synthesis and experimental evaluation a custom developed unmanned tri-tiltrotor is employed, equipped with rotor-tilting mechanisms which enable the direct actuation of its longitudinal dynamics, while retaining the standard body-pitching underactuated authority. An explicit model predictive control scheme relying on constrained multiparametric optimization is proposed for the dual-authority optimal control. The backbone of this scheme is a modeling representation that incorporates the separate internal dynamics of the two actuation principles and their interferences as they concurrently act on the free-flying vehicle body, while tractably representing their differentiated effects on the evolution of the longitudinal dynamics. This paper additionally presents the key implemented features that enable the autonomous operation of the employed tilt-rotor platform, in order to provide a reliable testbed for experimental evaluation. Finally, extensive experimental studies which conclusively validate this strategy's increased efficiency are demonstrated.