An experimental study of hierarchical intrusion detection for wireless industrial sensor networks

Wireless industrial sensor networks are necessary for industrial applications, so that wireless sensor nodes sense around themselves and detect anomaly events in the harsh industrial environments. Due to the harshness, anomaly events such as adversarial intrusions may result in harmful and disastrous situations for industrial applications but it is difficult to detect them over wireless medium. Intrusion detection is an essential requirement for security, but as far as we know, there have not been such studies for wireless industrial sensor networks in the literature. The previous intrusion detection methods proposed for wireless sensor networks consider networks rather in general senses and restrict capabilities to specific attacks only. In this paper, we first study intrusion detection for wireless industrial sensor networks, through various experiments and design of a hierarchical framework. We classify and select better methodologies against various intrusions. Subsequently, we find novel results on the previous methodologies. We also propose a new hierarchical framework for intrusion detection as well as data processing. Throughout the experiments on the proposed framework, we stress the significance of one-hop clustering, which was neglected in the previous studies. Finally, we construct required logical protocols in the hierarchical framework; hierarchical intrusion detection and prevention protocols.