Ensemble Deep Learning Classifier with Optimized Cluster Head Selection for NIDS in MANET

A MANET security is more fragile and susceptible to the environment due to the lack of a centralized environment for monitoring the behavior of individual nodes during communication in this type of network. Both local and global invaders are able to access the networks they target. In MANETs, where nodes can move in any direction and topology is constantly changing, node mobility and node energy are two critical optimization challenges. As a result, remote monitoring of node performance and behavior is employed by Network Intrusion Detection Systems (NIDSs) as a solution to cope with the problem of intrusion into these networks. The proposed method is used to develop a Cuttlefish Algorithm with Ensemble Deep Learning Classifier (CFA-EDL) for multi-attack intrusion detection. A clustering algorithm for MANET cluster head election is developed in this research by focusing on the challenges of mobility and energy. To select the cluster head, the CFA uses the EDL Classifier, while the EDL Classifier identifies several attacks. Multiple attacks are identified using EDL Classifier. Extensive testing in MATLAB and comparisons with other existing methods are included in the planned research. Attack detection, memory ingesting and computing time for classifying an intruder are some of the metrics used to evaluate the suggested method's performance. The results of the simulation show that the suggested strategy significantly reduces IDS traffic and memory ingesting while maintaining an attack detection rate in the shortest amount of time possible.