In recent years, Internet of Things (IoT) has received increased attention, and government and research institutions have actively cooperated in constructing IoT infrastructures. The discussions on IoT and home energy management systems have facilitated the development of appliance recognition technology, which assists users to effectively know electronic appliance usage, thus, further improving their decisions on power utilization. In consideration of the common power utilization habit of multiple electronic appliances, this study discussed multi-appliance recognition in the parallel state, i.e., appliance recognition for simultaneous use of multiple electronic appliances. This study proposed a design for an Appliance-oriented Home Energy Management (HEM) Platform based on the IoT architecture, where the service behaviors of electronic appliances are obtained from appliance detection, while energy management provides the service of integrating heterogeneous devices. The problem of the large data volume of current appliance recognition systems is solved by creating a database mechanism, appliance feature clustering, and a waveform recognition method. Different from the experimental environment of other recognition systems, parallel multi-appliance recognition and general user habits of power utilization were considered. Experiments in the routine habits of power utilization were conducted, with an average system precision rate of 92.07%, and an average single appliance recognition rate of 93.46%. The results proved that this study is highly feasible.
All Science Journal Classification (ASJC) codes
- Computer Networks and Communications