Machine learning based approach to detection of anomalous data from sensors in cyber-physical water supply systems
Alexey Meleshko, Vasily Alekseevich Desnitsky, Igor Vitalievich Kotenko · IOP Conference Series Materials Science and Engineering · 2020
Abstract The paper proposes an approach to detection of anomalous data from sensors in cyber-physical systems on an example of a water supply supply system. The approach is based on methods of machine learning and modeling of technical systems. The primary data for machine learning was obtained on the developed software/hardware prototype of the water supply system by using a number of microcontrollers, sensors and actuators. The experiments confirmed the feasibility of the proposed approach. Several machine learning methods from scikit-learn library of the Python programming language were tested. As a result of the experiments we identified a learning method and its parameters ensuring the highest accuracy of the abnormal situations recognition.