Wormhole attack detection method based on node pair similarity, shortest path hop count, and Bayesian classification algorithm in wireless sensor networks

Wireless sensor networks (WSNs), as an important component of the Internet of Things, have attracted much attention. However, they also face various security threats, among which wormhole attack is particularly notable. Wormhole attack causes severe harm to WSNs and is difficult to be detect. To effectively detect wormhole attack, this paper proposes a wormhole attack detection method based on node pair similarity, shortest path hop count, and Bayesian classification algorithm (WADM-NPS&SPHC&BC). WADM-NPS&SPHC&BC first screens out suspicious nodes by using the upper threshold of neighborhood growth rate and adds them to a suspicious node set. Then, it uses the similarity subthreshold to screen out suspicious node pairs from the suspicious node set and adds them to a suspicious node pair set. Finally, it detects node pairs under wormhole attack from the suspicious node pair set by using the upper threshold of path hop and the Bayesian classification algorithm. The simulation results show that after using the similarity subthreshold, WADM-NPS&SPHC&BC significantly reduces the computational overhead of calculating the shortest path hop count between exclusive neighbors of adjacent node pairs, thereby lowering the energy consumption of network nodes. Moreover, WADM-NPS&SPHC&BC shows greater advantages in terms of suspicious node coverage rate, number of fake links, and normal link false detection rate, compared with other similar detection methods.