Machine Learning‐Enabled Defending Model for Surveillance and Abnormal Behavior Detection Against Sinkhole Attacks in Mobile Networks

ABSTRACT Sinkhole attacks pose a serious threat to network security inherently, and how to detect and defend against them effectively is very important. In this paper, we study how to detect sinkhole attacks by three proposed algorithms. The first algorithm tests all possible paths in the network and abnormal traffic in mobile networks. It works in a pattern‐detecting manner and then configures multiple major nodes to explore all possible paths leaving each node as well as excluding paths that have detected abnormal behavior from the network and transportation area. The second algorithm is a method in which each network node independently determines its behavior and, when a particular pattern is detected, considers the node a sinkhole attacker and stops using the path. The third algorithm uses a basic machine learning model to learn data and then generates a hostile example to create a sinkhole. Also, it identifies and isolates nodes accurately. These algorithms can help increase network security and effectively defend against sinkhole attacks. This paper deals with the design and implementation of the proposed algorithm and its performance evaluation. In addition, the advantages and disadvantages of each algorithm are compared and analyzed to present various countermeasures against sinkhole attacks.