Control of chaos with minimal information transfer

Abstract

This paper studies set-invariance and stabilization of hyperbolic sets over rate-limited channels for discrete-time control systems. We first investigate structural and control-theoretic properties of hyperbolic sets, in particular such that arise by adding small control terms to uncontrolled systems admitting (classical) hyperbolic sets. Then we derive a lower bound on the invariance entropy of a hyperbolic set in terms of the difference between the unstable volume growth rate and the measure-theoretic fiber entropy of associated random dynamical systems. We also prove that our lower bound is tight in two extreme cases. Furthermore, we apply our techniques to the problem of local uniform stabilization to a hyperbolic set. Finally, we discuss an example built on the H\'enon horseshoe.