Dual mode predictive tracking of piecewise constant references for constrained linear systems (Q4800352)

The paper studies the tracking of a piecewise constant reference for a linear time-invariant (LTI) discrete-time system in the presence of state-control constraints and bounded disturbance inputs. The system under consideration is of the form \[ \begin{aligned} x(t+1)&=Ax(t)+Bu(t)+Dw(t),\\ y(t)&=Cx(t), \end{aligned}\eqno (1) \] where \(x(t)\in\mathbb R^n\), \(u(t)\in\mathbb R^m\), \(y(t)\in\mathbb R^p\), \(w(t)\in\mathbb R^q\), \(A,B,C,D\) are constant matrices and the pair \((A,B)\) is stabilizable. The plant (1) is subjected to the control and state constraints \[ u(t)\in\mathcal U\subset\mathbb R^m,\quad x(t)\in\mathcal X\subset\mathbb R^n,\quad t\geq 0, \eqno (2) \] the disturbance input satisfies \[ w(t)\in\mathcal W\subset\mathbb R^q,\quad t\geq 0, \eqno (3) \] where \(\mathcal W\) is a bounded set. The control objective is to make the output \(y(t)\) follow a given reference \(r(t)\) while satisfying constraints (2) for all disturbances (3). By using robust controlled invariant sets in the extended state-setpoint space, the paper gives a novel predictive tracking algorithm (named dual mode predictive tracker, DMPT) which yields robust constraint satisfaction and good tracking performance with a reasonable on-line computational burden. The algorithm operates in two different modes: as a standard predictive regulator whenever feasibility holds, or otherwise it attempts to recover feasibility as soon as this is lost due to setpoint changes. Simulation experiments demonstrate the good tracking properties of DMPT.