Intrusion Detection Based on k-Coverage in Mobile Sensor Networks With Empowered Intruders

Intrusion detection is one of the important applications of wireless sensor networks (WSNs). Prior research indicated that the barrier coverage method combinedwithmobile sensor networks (MSNs) can enhance the effectiveness of intrusion detection by mitigating coverage holes commonly appearing in stationary WSNs. However, the trajectories of moving sensors and moving intruders have not been investigated thoroughly, whereas the impact between two adjacent moving sensors and between a moving sensor and a moving intruder are still under-determined. In order to address these open problems, in this paper, we firstly discuss the virtual potential field between sensors as well as between sensors and intruders. We then propose to formulate the mobility pattern of sensor node using elastic collision model and that of intruder using point chargemodel. The point chargemodel describes an hitherto unexplored mobility pattern of empowered intruders, which are capable of acting upon the virtual repulsive forces from sensors in order to hide them away from being detected. With the aid of the two models developed, analytical expressions and simulation results demonstrate that our proposed design achieves a higher k-barrier coverage probability in intrusion detection compared to that of the conventional designs. It is also worth mentioning that these improvements are achieved with shorter average displacement distance and under much more challenging MSNs settings.