Investigation of Computational Intelligence Techniques for Intrusion Detection in Wireless Sensor Networks

Abstract

Wireless Sensor Networks (WSNs) have become a key technology for the IoT and despite the obvious benefits, challenges still exist regarding security. As more devices are connected to the internet, new cyber attacks are emerging which join well-known attacks to pose significant threats to the confidentiality, integrity and availability of data in WSNs. In this paper, we investigate two computational intelligence techniques for WSN intrusion detection. A back propagation neural network is compared with a support vector machine classifier. Using the NSL-KDD dataset, detection rates achieved by the two techniques for six cyber attacks are recorded. The results show that both techniques offer a high true positive rate and a low false positive rate, making both good options for intrusion detection. In addition, we further show the support vector machine classifiers suitability for anomaly detection, by demonstrating its ability to handle low sample sizes, whilst maintaining an acceptable FPR rate under the required threshold.