Enhance Privacy Protection by Reducing Information Exchanges in Multi-agent System Consensus

Frequent (agent-state) information exchanges for consensus in a multi-agent system (MAS) increases not only communication overhead but also the risk of sensitive information disclosure. This paper addresses the challenge of how to enhance data privacy protection by reducing information exchanges in MAS consensus based on event-trigger control (ETC). Unlike existing methods that inject uniform random noise into all agents and use trigger functions based on absolute state information, we propose a method that customizes noise injection according to agent importance with an adaptive trigger function to dynamically adjusts the communication frequency using the relative state information of agents, thereby reducing unnecessary transmissions while maintaining effective consensus. Experimental results verify the effectiveness of the proposed algorithm.