Support vector machine through detecting packet dropping misbehaving nodes in MANET

Abstract

Mobile ad-hoc network (MANET) is suffering from various attacks due to its infrastructure-less characteristics. Hence, MANET needs a variety of specific security methods to detect the false entrance of the misbehaving nodes. The network works well if the nodes are trusty and act rightly and cooperatively. In this paper, we are identifying and detecting the packet dropping nodes using the Support vector machine (SVM). A SVM is reactively used to classify the nodes in two different classes, either the normal or malicious nodes. SVM takes as an input the neighbor trust value, calculated with the data packets and control packets. Our technique is implemented with Ad-hoc on demand vector routing protocol (AODV). Our experimental results evaluated using the packet delivery ratio (PDR), End-To-End delay, Average throughput.