Real-Time Application of Deep Learning to Human Detection System for Smart Room Power Management

Population growth in residential, commercial and industry buildings have resulted in increasing demand for electrical power supply and energy. However, the limited available energy and power supply are characterized by wastage especially by the consumers. In addressing this problem, there is need to leverage on artificial intelligence (AI) for managing several energy or power sources and loads thereby saving more energy and cost. Thus, this paper presents a developed AI-based system named Intelligent Human-Sensing Power Management System (IHSPMS) that senses the presence of a human being to provide better monitoring and control of electrical appliances to save energy and cost. The IHSPMS provides smart room power management using digital web cameral as the sensing device and a microcontroller containing trained machine-learning (ML) model to detect humans in a room and appropriately triggers the relay to turn “ON” or “OFF” the room appliances. The system was tested with different scenarios and the results showed successful detection of human presence and absence in the room with switching “ON” and “OFF” of the appliances in each case as required. The proposed IHSPMS outperformed the non-AI based power management system in terms of energy efficiency.