Wicked node detection in wireless ad-hoc network by applying supervised learning

Abstract

A wireless ad-hoc network (WANET) is a decentralized network supported by wireless connections without a pre-existing architecture. However, the mobility of nodes is a defining characteristic of WANETs, and the speed with which nodes may act poses several security risks. As a result of these wicked nodes, more data packets are lost, which might cause a significant delay. Thus, it is very important to identify wicked nodes in WANET. This work provides a support vector machine approach for detecting (SVMD) wicked nodes in the internet of things. The number of characteristics is reduced using the linear correlation coefficient (LCC) technique. With the LCC technique, we can precisely measure the strength of the connection between any two nodes while clearing the field of irrelevant information. Further, the support vector machine (SVM) algorithm may identify the wicked nodes by analyzing metrics such as the packet received ratio, packet delay ratio, and remaining energy ratio. The next step is to reach a verdict in which the wicked nodes are punished by being rendered inoperable. The simulation results show that the network latency is minimized, and the chance of missing detection is decreased using this method in WANET.