Analysis of household power consumption data for social safety net services

Abstract

Much interest in lonely death and activity of daily living (ADL) monitoring services through analysis from household power consumption data is increasing. For this, anomaly detection and power disaggregation are needed, respectively. However, the existing technologies suffer from inaccuracy problem, so they are not widely used. In this study, activity perception-based anomaly detection and appliance activation profile-based disaggregation methods are newly presented to improve accuracy. According to the experimental results, the proposed methods showed better performance than the existing methods.