Tailor-Made Gaussian Distribution for Intrusion Detection in Wireless Sensor Networks

Intrusion detection is one of the primary applications in wireless sensor networks (WSNs) and has considerable importance in practice. Node deployment strategy plays a decisive role in determining the intrusion detection capability of a WSN. A uniform random deployment strategy fails to provide additional safety to sensitive areas within the network. However, a Gaussian deployment strategy provides additional safeguard to key areas within the network area. Moreover, both the deployment strategies suffer from energy-hole problem. In view of this, we consider energy-hole problem mitigating tailored-made Gaussian deployment strategy. Initially, we investigate the intrusion detection problem in a tailor-made Gaussian distributed network considering both single-sensing and multiple-sensing detection scenarios under a realistic probabilistic sensing model theoretically. Extensive simulation is performed mainly to validate the correctness of modeling and analysis. Effects of the different network parameters on detection probability are also observed in detail. Finally, the effectiveness of the proposed approach is confirmed by comparing with its counterpart Gaussian deployment strategy under the considered scenarios. The results clearly demonstrate our approach's dominance in terms of detection probability under various network parameters.