Enhancing security and efficiency in MANETs: a clustering-based approach with CGRUN and AGTO optimization for intrusion detection and path establishment

Mobile Ad-Hoc Networks (MANETs) represent dynamic, infrastructure-less wireless networks formed by self-configuring mobile devices, facing challenges in routing, security, and resource constraints. This study introduces the Convolutional Gated Recurrent Unit Network (CGRUN), a novel deep learning architecture tailored for efficient sequential data classification within MANETs. CGRUN amalgamates the strengths of Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs), offering a robust solution applicable across diverse classification tasks. In parallel, an innovative optimization model termed Artificial Gorilla Troops Optimizer (AGTO) is proposed, specifically designed to enhance weight reduction processes within the CGRUN framework. AGTO demonstrates promising capabilities in optimizing and refining the classification model, addressing the challenges of efficient resource utilization and network performance enhancement within MANETs. Moreover, this research presents a Python-based tool specifically engineered for analysing and processing the Wireless Sensor Network Dataset (WSN-DS). The tool achieves an impressive overall accuracy rate of 99.81%, surpassing existing solutions.