Design and Development of an IoT-Enabled Mobile Surveillance Car for Real-Time Video Monitoring

Abstract

The strategic change of fixed-camera into mobile surveillance is an urgent need in the existing Internet of Things (IoT) system to address the geometric limitation of the conventional fixed-camera system. This paper describes the design and construction of an ESP32-based mobile robot, 0.19m x 0.13m in size, as a dynamic alternative to the traditional monitoring systems. As opposed to computationally intensive local SLAM (Simultaneous Localization and Mapping) approaches, the architecture suggested requires a non-SLAM autonomous navigation approach that carries out sophisticated processing in the Amazon Web Services (AWS). The device has a multi sensor array which has four directional ultrasonic rangefinders and passive infrared (PIR) modules hence it has the complete spatial awareness. The software architecture is intentionally intensive to the cloud, which uses AWS IoT Core to transmit the data safely and AWS Recognition to perform facial analytics with high accuracy. The results show that this is a cost-effective and scalable solution, which removes the blind-spot problem of a fixed CCTV and also avoids the cost-prohibitive nature of LiDAR-equipped autonomous platforms.