The main subject of this paper is the sensing of network anomalies that span from harmless impedance changes at some network termination to more or less pronounced electrical faults, considering also cable degradation over time. In this paper, we present how to harvest information about such anomalies in distribution grids using high frequency signals spanning from few kHz to several MHz. Given the wide bandwidth considered, we rely on power line modems as network sensors. We firstly discuss the front-end architectures needed to perform the measurement and then introduce two algorithms to detect, classify and locate the different kinds of network anomalies listed above. Simulation results are finally presented. They validate the concept of sensing in smart grids using power line modems and show the efficiency of the proposed algorithms.
Fast Wireless Sensor Anomaly Detection based on Data Stream in Edge Computing Enabled Smart Greenhouse
Qufu Normal University（曲阜师范大学）；Weifang University of Science and Technology（潍坊科技学院）；Nanjing University of Science and Technology（南京理工大学）；Shandong First Medical University (Shandong Academy Of Medical Sciences)（山东第一医科大学（山东省医学科学院））
Samsung Electronics Company Ltd.（三星电子株式会社）；Korea Advanced Institute of Science and Technology（韩国科学技术院）；University of California San Diego（加利福尼亚大学圣迭戈分校）
A task of anomaly detection for a smart satellite Internet of things system
School of Mathematics, Hefei University of Technology（合肥工业大学数学学院）；School of Electrical and Automation Engineering, Hefei University of Technology（合肥工业大学电气与自动化工程学院）；CCNU Wollongong Joint Institute, Central China Normal University（华中师范大学伍伦贡联合研究院）
LEAD1.0: A Large-scale Annotated Dataset for Energy Anomaly Detection in Commercial Buildings
Singapore Management University（新加坡管理大学）
现代建筑物密集地配备了智能电表，这些电表定期生成大量的时间序列数据，每天产生数百万个数据点。这些数据可用于发现底层负载、推断其能源消耗模式、环境因素的相互依赖性以及建筑物的运营属性。此外，它使我们能够同时识别电力消耗曲线中存在的异常情况，这是朝着节约能源和实现全球可持续发展迈出的一大步。然而，到目前为止，缺乏大规模的标注能耗数据集阻碍了正在进行的异常检测研究。我们通过发布一个公开的ASHRAE Great Energy Predictor III数据集的注释版本来促进这一努力，该数据集包含跨越一年的1，413个智能电表时间序列。此外，我们在我们的数据集上对八种最先进的异常检测方法的性能进行了基准测试，并比较了它们的性能。
