Routing and Security Based Ad-hoc Networks Configuration for Identification of Attack Using Reinforcement Learning Approach

Abstract

In many fields of Research, wireless sensor networks have grown in popularity. Depending on the dangerous situation the networks fields, there are more chances to attack wireless sensor network. This research explains how to locate rogue nodes in the networks based on investigation and responses from each node in the networks. In generally a malicious node blocks the transmission of signal to other nodes. So, in order to increase network performance, avoid much traffic for the server to buffer, causing them to slow down and eventually stop. Hence to identify the fake node in wireless sensor network is main aim of this research through Network Simulator with the help of Reinforcement Learning based routing algorithms. Further, attack vectors allow hackers to exploit system’s vulnerabilities including manual elements. It is suggested that a workable security framework for the WSN (Wireless Sensor Network) used to estimate traffic or packet loss and throughput ratio in order to detect fault identity. So enhancing the networks performance is based on request-response mechanism of the nodes, which engages with features at various levels of the protocol's system, monitors and analyzes typical patterns and alerts node to ensure their dangerous activities cannot transmit across the network. Some of monitoring basic functions through routing algorithm and NS2 include overseeing server CPUs, paying attention to network traffic, identifying patterns in error rates, alerting you about slow pages and combing through your access logs to find out how long requests usually take.