CDC99-REG0545 Unconstrained Receding Horizon Control:

This paper addresses the issue of region of attraction of stabilizing receding horizon control strategies for nonlinear systems. This study builds on earlier results of the authors on obtaining stability guarantees for Receding Horizon methods by using an a priori obtained Control Lyapunov function (CLF) as terminal cost. It is shown that as the horizon length is increased, the region of attraction grows and approaches the region of attraction of the infinite horizon control strategy as the horizon length approaches infinity. The main contribution of these results is that a terminal state constraint is not required as some other available methods do. Using this approach, the receding horizon method can be thought of as a compromise between the controller obtained from the CLF and the infinite horizon control law. Since no state equality or inequality constraints are required, the computational burden is reduced drastically. An illustrative example is provided