Edge AI Enabled IoT Framework for Secure Smart Home Infrastructure
Pankaj Thakur, Shubham Goel, Emjee Puthooran
Abstract
The synergy of Edge Computing, Artificial Intelligence, and Internet of Things in smart city camera-based surveillance offers notable advantages in resource allocation. Traditional approaches involving constant video data streaming to central servers incur significant bandwidth and storage costs. Our contribution involves proposing a cost-effective, lightweight Edge AI Enabled IoT Framework for Secure Smart Home Infrastructure, utilising a Raspberry Pi single-board computer and the open-source software motion for camera surveillance. The motion program monitors video signals from various cameras and triggers specific actions upon detecting movement. The framework efficiently notifies the smart home owner via email and smartphone message when motion is detected. We integrated four advanced motion detection and alert notification methodologies, conducting a thorough evaluation that positioned our framework as superior to existing solutions. Our research showcases impressive accuracy rates of 91% and 85% in indoor and outdoor scenarios, with minimal average delays of 12.8 seconds for email alerts and 1.6 seconds for messages which is approximately 41.6 % less than the state of the art methodologies. This innovative integration not only elevates surveillance capabilities but also establishes a swift and reliable alert system, contributing significantly to the efficiency and security of smart home environments.