Introducing a New Benchmarked Dataset for Mechanical Stop Detection of Stepper Motors

Datasets are increasingly becoming an important research tool in most machine learning fields. In this paper, we address the lack of a publicly available dataset for mechanical stop detection of stepper motors. Various signals from 32 stepper motors are collected at different environmental conditions, to create an extensive dataset. The experimental setup, on which the dataset is based, is explained in detail and an overview of the structure of the dataset is given. Furthermore, benchmarking problems are established, discussed, and initial results, which are obtained using well-known standard machine learning methods, are presented. In addition to that, we identify the challenges associated with performing classification under the described environmental conditions of the recorded data.