Detection of Black and Grey Hole Attacks Using Hybrid Cat with PSO-Based Deep Learning Algorithm in MANET

Abstract

The newest example of wireless networks, known as mobile ad hoc networks (MANETs), offers some qualities, including a topology that can change dynamically, a baseless network, a range of transmission, a routing procedure, and reliability. In a black hole attack on a computer network, packets are deleted as opposed to being forwarded through a router. This often happens when a router has been corrupted by several circumstances. A routing attack called a "black hole" has the power to bring down an entire network. One of the most common types of assaults on MANETs is the Grey Hole Attack, in which a hostile node allows routing but prevents data transmission. MANET security is a top priority because they are far more susceptible to assaults than wired infrastructure. This study focused on detecting black and grey-hole attacks in MANET by using deep learning techniques. The forwarding ratio metric is used in the individual attack detection phase to distinguish between the defective and normal nodes. The encounter records are manipulated by malicious nodes in the collusion attack detection phase for escaping the detection process. The attacks are detected by using different deep learning techniques like Convolutional Neural Network (CNN) and a Long Short-Term Memory (LSTM) network. The parameter tuning operation is carried out by using the Hybrid Cat-Particle Swarm Optimization (HCPSO). The simulation results shown in our proposed system detect with better accuracy.