Resumen
This publication presents the development of an Industrial-Internet-of-Things device. The device is capable of completing several tasks, such as the acquisition of high-frequency measurement data and evaluating data via machine learning methods in an artificial intelligence application. The installed measurement technology generates data which is comparable to data generated by costly laboratory equipment, meaning that it can be used as a low-budget and open-source alternative. A workflow method has been designed that promotes experimental work and simplifies the effort required to implement artificial intelligence solutions. At the end of this paper, the results of the experiment, which aimed to collect measurement data, extract suitable features, and train artificial intelligence models, are presented. Techniques from vibration analysis were used for feature extraction, and concepts for the extrapolation and enhancement of data sets were investigated. The test results have proven that the development is comparable with high-end laboratory equipment. The created application has demonstrated sufficient accuracy in predictions, and the designed process can be used for arbitrary, artificial intelligence-based rapid prototyping.