Event Detection for Energy Consumption Monitoring

Abstract

The accurate detection of appliance state transitions in electrical signals is fundamental for numerous energy-conserving applications. We present an extensive overview and categorization of the current state in event detection on high-sampling-rate signals. Existing approaches are designed for specific environments and need to be tediously adapted for new ones. Thus, we propose an unsupervised, multi-environment event detector, outperforming four state-of-the-art algorithms on two heterogeneous public datasets.