Manual tracking control with amplitude and rate constrained actuators

Abstract

Tracking control in a receding horizon framework is considered and a novel, optimization-based, linear tracking control strategy-linear quadratic tracking (LQT) is devised. LQT yields a closed-form solution to optimization-based tracking of a dynamic reference signal r(t), in the face of both rate and amplitude actuator constraints; Actuator dynamics are included. Also, there is no inherent requirement for stability of the open-loop plant. At the same time, full state feedback is assumed. The detrimental effects of the actuator constraints are mitigated by giving the control system a nonlinearly modified, feasible reference signal r'(t), as required to prevent downstream actuator saturations from occurring. Thus, the controller generated signal never infringes on the saturation bounds, and windup is precluded. The analytic solution to the linear, unconstrained tracking problem meets the small signal performance specs and is stable. The proposed piecewise linear closed form solution to the constrained tracking problem yields good responses to large inputs and at the same time requires modest online computation, and hence is implementable in real-time.