PROJECT TITLE :

Traffic Anomaly Detection in Wireless Sensor Networks Based on Principal Component Analysis and Deep Convolution Neural Network

ABSTRACT:

Because of the proliferation of wireless networks, wireless sensor networks (WSNs) have developed very quickly. However, the flexibility and ease of deployment of WSNs have led to an increase in the number of security concerns; consequently, it is essential to conduct research on the prevention of network intrusions for WSNs. The denial of service attack, also known as DoS, is a common form of network attack that aims to bring down the network it is attacking. DoS attacks on devices in WSNs, which typically have very few resources, would be catastrophic. The purpose of this paper is to propose a method for the detection of DoS traffic anomalies in WSNs that is based on principal component analysis (PCA) and a deep convolution neural network (DCNN). The paper bases its proposal on the susceptibility of WSNs to attacks as well as the limited storage space of their devices. When compared to the conventional Deep Learning structure, the proposed model has a lightweight structure and a more effective capability for Feature Extraction. As a result, it is able to detect network abnormal traffic in WSNs devices that have a limited capacity for storage while still being effective. Verifying the classification results of the model is done with the help of receiver operating characteristic (ROC) curves, a variety of classification metrics, and confusion matrices. This is done so that the efficacy of the proposed model can be ensured. The proposed model, despite its small model size, has been shown to perform better than other mainstream abnormal traffic detection models in terms of its classification effect in the context of experimental comparison.