Local Monitor Implementation for Decentralized Intrusion Detection in Secure Multi-Agent Systems

Abstract

This paper focuses on the detection of misbehaving agents within a group of mobile robots. A novel approach to automatically synthesize a decentralized Intrusion Detection System (IDS) as well as an efficient implementation of local monitors are presented. In our scenario, agents perform possibly different independent tasks, but cooperate to guarantee the entire system's safety. Indeed, agents plan their next actions by following a set of logic rules which is shared among them. Such rules are decentralized, i.e. they depend only on configurations of neighboring agents. However, some agents may not be acting according to this cooperation protocol, due to spontaneous failure or tampering. To detect such misbehaviors, we propose a solution where each agent runs a local monitor that uses only locally available information. In this paper, we present an implementation of such monitors by which events occurred to a target-agent can be estimated for any combination of neighborhood and observable space. Validity of the proposed implementation is shown through simulation.