Preventing Black Hole Attack and Safe Data Transfer in MANET Using Neural Network and RVM Classifier

Abstract

Mobile ad-hoc network is a developing area of interests, which utilized in a wide range of applications. A MANET is a self-sufficient group of mobile users that impart over data transmission of constrained with wireless connections. Since the nodes are versatile, the network topology may change quickly and unusually over time. The mobile nodes are self-configured network that were formed in anytime, anywhere without interfering centralized management or infrastructure. While transmitting the data the security attacks were formed because of open centralized management. Main cause of attack is black hole attack at most of the time. Black hole is one type of attack which makes the packets loss or dropping the packets and then sends the acknowledgement as the packet send, hence it is a serious malicious activity.  As a user point view the packets forward to the target/node, but actually the packets are not sent properly. So to avoid this problem proposes the Path Identification of Black Hole Check (PIBHC) using neural network. Which classifies the attacker path node or by using the Relevance Vector Machine (RVM) to proceed further data transmission.