Multi-view consistent generative adversarial network for enhancing intrusion detection with prevention systems in mobile ad hoc networks against security attacks

Improving security in Mobile Ad hoc Networks (MANETs) requires an effective intrusion detection and prevention scheme that addresses some research issues, such as energy efficacy, delay, detection rate, false positive rate. However, many existing solutions have faced challenges in achieving accurate detection rates while minimizing execution time and energy consumption. In this manuscript, a Multi-View Consistent Generative Adversarial Network for Enhancing Intrusion Detection with Prevention Systems in MANET Against Security Attacks (IDPS-MANET-MVCGAN) is proposed. Initially, the mobile users are registered in Trusted Authority under One Way Hash Chain Function. The intrusion detection is executed using four entities. In the packet analyzer, it is verified regarding if any attack is identified or not. The implementation is done in Type 2 Fuzzy Controller that takes data through packet header. The collected data is fed to improved splice Kalman filtering for data normalization. Then it is supplied to the feature extraction using Multi-Scale Ternary Pattern Mutual Information to extract the optimum set of features for packets classifcation. During classifcation, Multi-View Consistent Generative Adversarial Network (MVCGAN) is used for packets classification as DoS, Probe, U2R, R2L, Normal. To improve the accuracy of the method, Fire hawk optimization algorithm (FHOA) is used. The proposed IDPS-MANET-MVCGAN method attains 13.88 %, 25.75 %, 16.16 % better accuracy when compared with the existing models: Adaptive Marine Predator Optimization Algorithm Deep Supervised Learning Classification dependent Intrusion Detection Scheme for MANET Security (IDPS-MANET-DSLC), An Intrusion Detection Scheme utilizing Exponential Henry Gas Solubility Optimization based Deep Neural Fuzzy Network in MANET (IDPS-MANET-DNFN) and Adaptive Activation Functions along Deep Kronecker Neural Network optimized with Bear Smell Search Algorithm to prevent Cyber security attacks in MANET (IDPS-MANET-ADKNN) respectively.