CoCoIDS: A Collaborative Intrusion Detection System for IoT based on Co-evolution

The vulnerability of the Routing Protocol for Low-Power and Lossy Networks (RPL) to insider attackers highlights the importance of detecting and responding to malicious attempts to ensure uninterrupted protocol operation. Researchers are generally developing collaborative Intrusion Detection Systems (IDS) to increase detection performance, yet achieving a balance between effectiveness and efficiency remains challenging, often leading to compromises. Here, we introduce CoCoIDS, a Collaborative Co-evolution Based Intrusion Detection System for IoT. Unlike existing approaches, our solution prioritizes both effectiveness and efficiency, including communication cost and resource consumption in collaborative nodes. Through a cooperative co-evolution strategy based on multi-objective optimization, our approach simultaneously optimizes the detection algorithm and collaborator node selection. Minimizing the number of collaborator nodes is especially important in a lossy environment, where communication disruptions can block collaborative efforts. The results show a balanced trade-off between accuracy and the number of collaborator nodes. Additionally, the study thoroughly analyzes selected collaborators based on their locations and targeted attack types.