Detection of blackhole and wormhole attacks in WSN enabled by optimal feature selection using self-adaptive multi-verse optimiser with deep learning

The enormous growth of data that are transmitted through diverse devices and communication protocols have raised critical security concerns. This, in turn, has amplified the significance of introducing more highly developed intrusion detection systems (IDSs). The main intent of this paper is to introduce new intrusion detection systems (IDSs) mainly focusing on blackhole and wormhole attacks in WSN with the aid of improved deep learning algorithms. To capture the optimal features with unique information, optimal feature selection is introduced with the assistance of a new metaheuristic algorithm called self adaptive-multi-verse optimisation (SA-MVO). Finally, optimally selected features are subjected to a deep learning algorithm termed as deep belief network (DBN). In the detection side, the same proposed SA-MVO is used to improve the DBN by optimising the number of hidden neurons. The results demonstrate that the proposed approach improves the detection probability when compared to conventional methods by analysing diverse performance measures.