Non-Intrusive Appliance Monitoring Now: Effective Data, Generative Modelling and LETE
Monitoring energy consumption at appliance level is a necessary condition for many energy efficient applications. In this paper, the authors have identified Non-Intrusive Load Monitoring (NILM) as a transitional technology, and competing NILM methods must be evaluated not only on performance, but also economical feasibility and user usability of now. Since, in the event of maturing an IoT infrastructure, there would be no need for NILM. In this paper, they proposed a NILM for now, a complete solution with novel features: pre-processing for effective data, thus more accurate and faster training for user, generative modelling for appliance configuration states, thus without the need for exhaustive training, and LETE algorithm for a simple stateful enhancement.