A Smart Socket for Real-Time Nonintrusive Load Monitoring
Wu Zhao, Chao Wang, Linyun Xiong, Ruiheng Li, Tao Wu, Huaiqing Zhang
Abstract
In the field of nonintrusive load monitoring (NILM), scientists across various institutions have published extensive research on both the hardware and the load disaggregation algorithm. However, there are few publicly available hardware design schemes for real-time load disaggregation. This increases the barrier to entry into the field. In this article, a novel smart socket for load disaggregation by using a set of easily accessible sensors is presented, aiming to enable more researchers to actively participate in NILM quickly. In addition, we combine a time-efficient automatic state detection algorithm and a factorial hidden Markov model to achieve the goal of real-time load disaggregation. The proposed method also alleviates the problem of requiring extra information that broadly exists in the traditional hidden Markov model-based load disaggregation algorithms. Experiments on a public dataset as well as a dataset collected from our laboratory are conducted, and the results confirm the efficiency and effectiveness of the proposed smart socket.