Low-cost prototype for bearing failure detection using Tiny ML through vibration analysis.

Journal: HardwareX
Published Date:

Abstract

The document presents a low-cost, open-source device designed to facilitate the learning of technologies like artificial intelligence in embedded systems through vibration analysis. It also aims to enhance students' skills by introducing industrial challenges into the classroom via a scaled-down prototype. This study analyzes the vibrations generated by bearings to classify, using Artificial Intelligence (AI), whether they are defective. The device integrates electronic, mechanical, and software components, leveraging online technologies and platforms like Arduino to support hands-on learning. The document provides detailed instructions on the components used, circuit connections, step-by-step construction, and implementation, allowing replication of the prototype. This device fosters the development of STEM skills, promotes the application of AI and TinyML in real-world contexts, and enriches educational programs by encouraging interdisciplinary learning.

Authors

  • Andres Felipe Cotrino Herrera
    School of Engineering and Basic Sciences, Universidad Autónoma de Occidente, Cali, Colombia.
  • Jesús Alfonso López Sotelo
    School of Engineering and Basic Sciences, Universidad Autónoma de Occidente, Cali, Colombia.
  • Juan Carlos Blandón Andrade
    Systems and Telecommunications Engineering Program, Universidad Católica de Pereira, Pereira, Colombia.
  • Alonso Toro Lazo
    Systems and Telecommunications Engineering Program, Universidad Católica de Pereira, Pereira, Colombia.

Keywords

No keywords available for this article.