Detection of Black Hole Attack Using Honeypot Agent-Based Scheme with Deep Learning Technique on MANET

Abstract

Mobile Ad-Hoc Networks (MANETs) due to their reconfigurable nature are being integrated into new and futuristic knowledge such as Internet of Things (IoT), cloud, reconfigurable networks, etc. To attain such credibility of integration, the routing protocols associated with these mobile nodes have to connect, perform and facilitate routing that offers a high level of security and resistance to all possible threats and security issues that may emanate in the network. One of the solutions used to maintain network security is intrusion detection systems (IDSs). This article primarily emphasis on the network's susceptibility to a suction assault known as a black hole attack. The investigations about the employment of intelligent agents called Honeypot Agent-based detection scheme (HPAS) with Long-Short Term Memory (LSTM) in identifying such assaults. Hence, the proposed method is named HPAS-LSTM, where honeypots are roaming virtual software managers that create Route Request (RREQ) packets to attract and entrap black hole attackers. Extensive model results utilizing the ns-2 simulator are used to demonstrate the presence of the suggested detection technique. The simulation outcomes demonstrate that the suggested technique outperforms current black hole detection methods in terms of throughput (TH), packet loss rate (PLR), packet delivery ratio (PDR), and total network delay (TND).