Abstract

Mobile Ad-hoc Networks (MANETs) face significant challenges in maintaining connectivity and security due to their dynamic and decentralized nature. Passive attacks, such as eavesdropping and traffic analysis, pose a critical threat to MANET. This study proposes a novel algorithm, Radial ResNet, tailored for the classification of passive attacks in MANETs. The algorithm integrates radial basis function networks with residual networks (ResNet) to enhance classification accuracy and efficiency. Experimental results demonstrate the effectiveness of Radial ResNet, achieving an average classification accuracy of 95.2% across various passive attack scenarios. Moreover, the algorithm exhibits improved computational efficiency compared to traditional methods, reducing processing overhead by 30%. The proposed approach not only enhances security by accurately identifying passive attacks but also optimizes network performance by mitigating resource-intensive computations.