Robust time-optimal control of constrained linear systems (Q1388103)

The authors consider the discrete-time dynamic system \[ x[k+1]=Fx[k]+gu[k]+w[k], \tag \(*\) \] with the state vector \(x[k] \in R^{n}\), the control function \(u[k] \in R^{m}\), and the perturbation \(w[k] \in R^{n}\), \(m < n\). The constraints on the control function and the state vector are given as: \[ u[k] \in \Omega, \;x[k] \in E, \] where \(\Omega \in R^{m}\) is convex and compact, \(E \subset R^{n}\) is convex and closed. One also assumes that the perturbation \(w\) is a constrained function for each \(k \in N\), where \(W\) is a convex and compact set with the origin in its interior. The main idea of the paper is the synthesis of a nonlinear controller of the form \(u[k]=h(x[k])\) such that the closed-loop system \[ x[k+1]=Fx[k]+Gh(x[k])+w[k] \] is independent from the perturbations \(w\) and moreover is able: a) to transfer the dynamic object from a starting state to a target set \(X_{0} \subset E\) in minimum time, b) to stop the state continuously in the target set \(X_{0}\). Some examples are included.