Distance Vector-Hop (DV-Hop) and Differential Evolution (DE)-Based Interception Strategy for detecting Cross Border Infiltration in Underground Tunnel

Abstract

Securing the external border of a nation through potential surveillance is considered being highly essential, as the infiltration of trespassers and terrorists is considered to influence the harmony and peace of a nation. In this context, intensive human patrolling is required for safeguarding longer border areas. In specific, detecting the infiltration of terrorists in the underground tunnel is yet another challenge. Wireless sensor networks deployment is necessary for safeguarding the external borders through real-time monitoring with maximised accuracy and detection efficiency. Most of the existing approaches failed in adopting an optimal sensor node deployment strategy that prevents maximised overlapping of sensor nodes' coverage and struggled to localise the intruders with maximised accuracy and minimised error rate of positioning. In this paper, Distance Vector-Hop (DV-Hop) and Differential Evolution (DE)-based Interception Strategy (DV-Hop-DE-IS) are proposed for accurate detection of cross border infiltration in the underground tunnel. This proposed DV-Hop-DE-IS adopted the disk model of optimal sensor node deployment for concentrating on the sensing area of the underground tunnel despite overlapping regions realised during the coverage process. This proposed DV-Hop-DE-IS includes the merits of converting the discrete values of hop count into a highly accurate continuous value depending on the information received from the number of shared one-hop nodes that exist between neighbouring nodes. This problem of intruder detection is planned as the minimum optimisation problem that could be optimally solved through the utilisation of the Differential Evolution algorithm with maximised efficiency. The simulation results of the DV-Hop-DE-IS confirmed a better detection rate of 6.84 per cent, improved accuracy of 11.28 per cent with a reduced false-positive rate of 8.28 per cent, compared to the benchmarked intruder detection approaches.

Description