Resilient Consensus with Mobile Detectors Against Malicious Attacks

Abstract

This paper investigates the problem of resilient consensus under malicious attacks for multiagent systems. Compared with most of existing works, a more general attack model is considered, where malicious agents can neighbor and collude with each other and the number of tolerable attacks is not limited by the network connectivity, which makes the problem more challenging. To solve this problem, we exploit the mobile agents as the detectors, and design the resilient consensus algorithm with mobile detectors (MRCA). MRCA includes basic average iteration process and the detection algorithms of both static and mobile agents. We prove that under MRCA, malicious agents will be detected by mobile agents and isolated by normal agents in finite time in expectation and thus the convergence can be achieved. We also show that MRCA can tolerate neighboring malicious attacks and the number of tolerable attacks is not limited by the network connectivity. Simulations are conducted to demonstrate the effectiveness of MRCA and validate the correctness of the theoretical results.