Sensor Integration and Facial Recognition Deployment in a Smart Home System

Abstract

Smart homes are systems that integrate sensing and control system, home appliances, and networking devices that can be accessed, authorised, and monitored digitally by the occupants. A smart home environment needs to address three significant aspects to be effective—occupant safety and comfort, energy efficiency, and convenience. This paper aims to provide a methodological framework for a smart home system (SHS) with integrated sensors and facial recognition system (FRS). The integrated sensor system consists of gas and fire sensors to make the smart home accident-free, turbidity sensors to monitor the quality of water consumed by the resident, and electric current sensors to detect anomalies in home appliances. It also deploys facial recognition algorithms such as HAAR-Cascade algorithm and Local Binary Pattern Histogram algorithm to enhance the SHS’s security. This proposed SHS is monitored in real-time by the occupant via a bot designed on the Telegram app. The Sensor Integration framework is developed by following a detailed iterative design process. The Facial Recognition System (FRS) is accomplished through a five-step development process, which is tested using a preliminary dataset of eight people and returns an accuracy of 87.5%. Testing in real-time conditions and developing a scalable model for multipurpose buildings are included in the future scope.