Intruder Detection System Through Walking Pattern Analysis for Home Security
Ravi Peiris, R. Tharmikka, Vijitha Rohana Herath, Maheshi Buddhinee Dissanayake, U. Sudheera Navaratne · 2021
Modern home automation systems have home security-enhancing features such as face detecting camera systems, and fire alarm systems. In this paper, a novel home security system is proposed which can detect intruders by analyzing walking patterns and updating the owner immediately. The system architecture adopts the concept of the Internet of Things (IoT), providing a network and user-friendly system, which supports simple expansion through plug-and-play devices. The paper first presents a simple hardware prototype of an IoT device to detect the walking patterns of home dwellers. Then it presents a Machine Learning based solution to analyze the walking patterns to detect intruders. The proposed prototype has shown promising performance in five machine learning algorithms tested with 96% of average performance accuracy in detecting intruders