Research Article
Embedded Machine Learning for Machine Condition Monitoring
@INPROCEEDINGS{10.1007/978-3-030-78459-1_16, author={Michael Grethler and Marin B. Marinov and Vesa Klumpp}, title={Embedded Machine Learning for Machine Condition Monitoring}, proceedings={Future Access Enablers for Ubiquitous and Intelligent Infrastructures. 5th EAI International Conference, FABULOUS 2021, Virtual Event, May 6--7, 2021, Proceedings}, proceedings_a={FABULOUS}, year={2021}, month={6}, keywords={Embedded machine learning Microcontroller Intelligent sensor Industrie40 IoT AI}, doi={10.1007/978-3-030-78459-1_16} }- Michael Grethler
Marin B. Marinov
Vesa Klumpp
Year: 2021
Embedded Machine Learning for Machine Condition Monitoring
FABULOUS
Springer
DOI: 10.1007/978-3-030-78459-1_16
Abstract
With the application of a new generation of information technology in the field of manufacturing and the deep integration of computer technology and manufacturing, industrial production is moving towards intellectualization and networking . Because the current production system cannot fully exploit the value of industrial data and the existence of information islands in the production process, this paper presents a study on the development, testing, and evaluation of a machine learning process that can be run on low-cost standard microcontrollers with limited computing and memory resources. This paper first analyzes the basic idea of whether it is possible to develop software for intelligent sensors whose algorithms run on microcontrollers. At the same time, it is considered whether the training and the adaptation of the model parameters can be done on the microcontroller to enable an online adaptation of the machine to be monitored. The goal is a closed system that does not need a backend and the storage of large amounts of data is not necessary.
