Abstract:
The rapid growth in the Internet of Things (IoT) has led to an increased risk of security attacks. An increase in network traffic has attracted cyber criminals and hackers to inject more network attacks into IoTs. The most common type of RPL-based network attack on an IoT network is a wormhole attack, which can have devastating effects on network performance and reliability.
As traditional security approaches like cryptographic protocols, Distance-based methods, and Signal strength-based methods may not be effective against wormhole attacks due to their basic level of security and due to the dynamic nature of the IoT network. In this paper, we propose a machine-learning approach for detecting wormhole attacks in IoT networks. In our approach, we used a new dataset that was generated in the Cooja simulator to train and test a binary classifier that can accurately distinguish between normal network traffic and wormhole attack traffic. A wormhole is a complex type of network attack that depends on the multiple types of features. So instead of using the limited type of features, we have used some additional features compared to the ones already used by the researchers. After preprocessing the dataset, we trained and tested using different classifiers using hit and trial method, but among of them Artificial Neural Network (ANN), Ridge classifier, Deep Belief network, and (RBFN) classifier give the best results. Our results demonstrate that the proposed approach achieves high accuracy in detecting wormhole attacks, making it a promising solution for enhancing the security of IoT networks.
Keywords : Artificial Neural Network, Radial basis function, Deep belief network
