Artificial Intelligence-Based Wireless Sensor Network Model for Intrusion Detection and Firearms Image Detection in the Conflict Zone

Abstract

The convergence of wireless sensor networks (WSN) and artificial intelligence (AI) for gathering security information about terrorism patterns movement in war zones renders advantages. It improves the efficiency of sophisticated machinery in combatting terrorists, which is more resilient than humans at the front line through autonomous surveillance. However, the detection of this terrorist pattern movement and human-conceived weapons with a traditional approach through the deployment of gallant soldiers and other anti-terrorist personnel to the conflict zone is more challenging. Which usually results in mass causalities, fatalities, and machinery destruction by ambushing. So, this research aims to develop an architecture for intrusion detection movement patterns, and firearm detection models in the war zone through the deployment of wireless sensor nodes, and autonomous surveillance camera systems. The imagery of human-conceived weapons is collected and experimented with the YOLOv5 model for object detection, and classification using a deep convolutional neural network (DCNN) embedded with the YOLOv5 platform. The average detection accuracy results obtained for the simulation of attacker detection in a spanning tree network-based wireless sensing model are 94.85%, 95.10%, 96.58%, 93.57%, 95.26%, and 97.17% respectively. Also, the result obtained for the detection accuracy of firearms identification is 100%, with a processing time of 0.875 seconds.