Nonlinear Control Algorithm for Systems with Convex Polytope Bounded Nonlinearities

Abstract

This paper describes a technique for controlling nonlinear systems. It is assumed that the nonlinearity takes values in a convex polytope, the control appears linearly, and the system can be discretized in time. The technique requires the solution of a finite number of linear feasibility (programming) problems and reconstructs the nonlinear control from these solutions. Several examples are provided to illustrate the technique and results are compared to feedback linearization.