Efficient Physical Intrusion Detection in Internet of Things: A Node Deployment Approach

In Internet of Things (IoT), intrusion detection plays an important role in many applications for detecting malicious intruders. The intruder can be, an unexpected physically moving entity, invading an area under surveillance, or an adversary in a battlefield. Node deployment strategy plays a crucial role in determining the intrusion detection capability of an IoT network. With uniform deployment, the detection probability is the same for any location in the network area. Nevertheless, different applications may need diverse levels of detection probability at key areas within the network. For example, a battlefield surveillance application needs improved detection probability around the headquarter. On the contrary, a Gaussian deployment strategy provides improved detection probability to the key areas due to differentiated node density. However, it is neither energy-efficient nor provides a quick detection of the physical intruder. In this work, we introduce a novel deployment strategy to overcome the above said limitations of both uniform and Gaussian deployments for energy-efficient and quick detection. Initially, we investigate the problem of physical intrusion detection in our introduced deployment strategy considering a realistic sensing model. Furthermore, we examine the effects of different network parameters on the detection probability in details. We also derive the relationship between the different network parameters and connectivity to ensure fast detection. We perform exhaustive experiments on real datasets, primarily, in order to validate the correctness of modeling and analyses. Next, we examine the effects of different network parameters on the detection probability. The results clearly demonstrate that our approach improves the detection probability by more than 25% when compared to two well-known deployment strategies under various network parameters.