Joonsu Ryu’s research while affiliated with Sogang University and other places

The proliferation of wireless mobile devices has resulted in mobile ad hoc network (MANET) technologies, which enable the formation of networks without infrastructure assistance, gaining increased attention. The advancements in these technologies have also resulted in the variety of attacks targeting them becoming more diversified. Particularly, owing to their inability to ensure node reliability, MANETs are vulnerable to routing attacks such as wormhole attacks. Since wormhole attacks often do not directly damage networks, detecting them can be challenging. In response, we propose a novel multiple verification-based wormhole attack detection method that leverages the characteristics of such attacks. The proposed method measures the credit of each node based on a trust system. The trust levels of suspicious nodes are reduced during routing; those with trust levels below a certain threshold are considered malicious. This trust system was implemented using reinforcement learning, which improves the accuracy of the system over time. Simulation experiments in which the proposed method was applied to existing routing methods in a densely populated environment were conducted; the rate of traffic passing through paths with malicious nodes was significantly reduced.