Feedback Control of Border Collision Bifurcations in Two-Dimensional Discrete-Time Systems

Abstract

The feedback control of border collision bifurcations is consideredfor two-dimensional discrete-time systems. These are bifurcations that can occur when a fixed point of a piecewise smooth system crosses the border between two regions of smooth operation. The goal of the control effort is to modify the bifurcation so that the bifurcated steady state is locally unique and locally attracting. In this way, the system's local behavior is ensured to remain stable and close to the original operating condition. This is in the same spirit as local bifurcation control results for smooth systems, although the presence of a border complicates the bifurcation picture considerably. Indeed, a full classification of border collision bifurcations isn't available, so this paper focuses on the more desirable (from a dynamical behavior viewpoint) cases for which the theory is complete. The needed results from the analysis of border collision bifurcations are succinctly summarized. The control design is found to lead to systems of linear inequalities. Any feedback gains that satisfy these inequalities is then guaranteed to solve the bifurcation control problem. The results are applied to an example to illustrate the ideas.