A model-predictive satisficing approach to a nonlinear tracking problem

Abstract

In this paper we use the recently introduced concept of satisficing decision theory in conjunction with a receding horizon optimization technique to achieve suitable tracking for a nonholonomic robot system. The satisficing approach creates a family of "universal formulas" parameterized by two functions. A model predictive scheme is employed to generate these two functions in a way that minimizes the quadratic cost at the next time step. By always choosing an element of the satisficing set, global stability is guaranteed.