Effective detection of a mobile intruder in a partially connected wireless sensor networks

Abstract

For a wireless sensor network (WSN) to operate successfully in an application of detecting a mobile intruder, the sensor's sensing of the intruder and the communication between the tasking sensor(s) and the base station should be jointly considered for effective intrusion detection. Most related work either treats the two tasks separately or considers both tasks for general-purpose WSN applications in terms of sensing coverage and network connectivity. This work instead investigates the effective intrusion detection problem in a partially connected random WSN from modeling, analysis, and simulation perspectives by integrating the K-sensing and communication tasks. Upper and lower bounds of effective K-sensing intrusion detection probability are mathematically formulated and theoretically derived. Monte-Carlo Simulations are conducted and outcomes are shown to support the theoretical analysis.