Porting and Optimization of Electric Power Consumers Anomaly Behaviour Monitoring on ARM Platform

With the development of smart grid, finer grid monitoring becomes possible. We propose a framework for abnormal behavior monitoring of power users, and carry out targeted deployment and optimization on the ARM platform. In order to align the anomaly detection performance of the model with that of humans, we designed a human-in-the-loop pipeline to achieve more accurate anomaly detection results with a small number of expert annotations. We validate the effectiveness of the proposed method on the electricity data of a region in China.