Enhanced Secure Routing in MANET using Collaborative Machine Learning Approach

Abstract

In Mobile Ad hoc Networks (MANET), routing security is critical. Real-world communication development aims to allow decentralized applications between nodes in dynamic topology situations. Intermediary nodes handle communication among nodes in a MANET. The interim node may often operate as just a fool node by executing any aberrant functionality. As a result, we must safeguard the suitable terminals. Even though the spammers assault the networks, the sender node must locate other intermediary nodes to deliver the payload to an endpoint. It is a severe problem with the MANET since the start node has no idea of a neighbor to choose in order to route. To fix this issue, we employ the novel approach, i.e., K-Dimensional (KD-Tree) with the KNN (K-Nearest Neighbours) approach. This technique uses to locate and safeguard intermediary nodes till they arrive at their end destination. We show the usefulness of the suggested technique by using the KD-KNN to identify compromised faulty nodes that perform assaults such as direct or indirect assaults and discover the closest safe nodes. The performance graph and packet drop are both executed in the resultant graph. The outcomes of research experiments suggest that our technique detects faulty nodes, locates the safest nearby node, and has a reasonable burden.