Adversary Control Strategies for Discrete-Time Systems

This paper addresses the concept of adversary control for a single-input dynamical system. The system evolves in the discrete-time domain and is subject to both state and input hard constraints. A control law guarantees positive invariance and contractivity of a bounded, convex, polyhedral set with respect to the system as well as asymptotic stability of the origin with maximum convergence rate. Periodically, an adversary controller succeeds in gaining control of the system and sends false control commands attempting to drive the state vector outside of the polyhedral set at the maximum admissible rate. An enhanced adversary policy which takes measurement errors into account is also developed. Simulation studies highlight the results of this sequential, non-cooperative control game.