Identification for Systems with Binary Data against Piecewise Constant DoS Attacks: A Game Learning Approach

Abstract

This paper proposes a game learning approach for system identification with binary data against piecewise constant DoS (Denial-of-Service) attacks. First, the game model of attack and defense is established, and the strategy sets and payoff functions of the attacker and the defender are given. Then, aiming at the piecewise constant DoS attack, a game learning rule is designed for the defender. Based on the rule, an attack strategy estimation algorithm and a system parameter estimation algorithm are constructed, and their performances are analyzed in a given stage. Finally, the rationality of the theoretical results is verified by numerical simulations.