Edge-based Energy Management for Smart Homes

With the penetration of Internet of Things (IoT) paradigm into household scenarios, more and more smart appliances are deployed to improve living comfort of residents. However, the rising electricity cost of these smart appliances become a critical challenge since existing home energy management systems (HEMS) are unable to effectively reduce electricity bills, but drive high cost of capital investment on the establishment of such infrastructure. In this work, we propose an edge-based energy management framework, which enables low electricity cost and saves budgets on building up the infrastructure. We employ a load shifting technique on an edge device, and solar energy is integrated into household energy system as a cost-effective solution. We propose an optimal scheduling strategy to schedule operation time of each appliance for achieving minimum electricity cost. This strategy basically requires to fully consider user preference on each appliance which can be derived from appliance-level energy consumption. In order to lower capital investment on equipment as sensors and smart meters, we use a location-awareness non-intrusive load monitoring (NILM) algorithm to achieve energy disaggregation at different times during a day. To validate our proposal, we implemented a prototype system with real smart appliances and Raspberry Pi. The experiment results demonstrated that electricity cost is significantly decreased (by 82.3%) when our proposed framework was employed compared to the cases without using our proposal.