Intrusion detection system using voting based neural network

Abstract

Remote sensor networks are increasingly being used in a variety of applications, including security and surveillance, control and support of complex frameworks, and fine-grained monitoring of indoor and outdoor variables. Remote sensor networks are rendered completely defenceless in the event of an attack. Because the portable hubs are distributed indiscriminately and there are no actual obstacles for the enemy, they can be easily captured, and assaults can come from any direction and target any hub. As a result, remote sensor network (WSN) security is the most difficult for this sort of industry. Intrusion Detection Systems (IDSs) can help detect and prevent security breaches.For data and correspondence innovation, an interruption discovery component is regarded as a major source of security. Due to particular constraints, such as asset-required devices, hubs with limited memory and battery capacity, and explicit convention stacks, traditional interruption location solutions should be altered and improved for usage in the Internet of Things. We present a lightweight attack detection technique based on a controlled AI-based IDS for identifying an adversary attempting to inject superfluous data into a network organisation in this study.