Scholarlink

This paper presents a systematic approach to data acquisition and analysis for estimating the 2D position of a permanent magnet using neural networks. In a step-by-step manner, the data were collected using a functional prototype, and the training was analyzed. Multiple neural network architectures were evaluated, and the best possible results were selected. This most effective network was exported as API and subjected to further real-world tests, and the results are presented. This research was conducted in an effort to create a sensor that can measure contact forces for robotic hands and grippers by measuring the change of the magnetic field of magnets suspended in a flexible substrate. The idea is that the methodology developed for determining the position of the magnet can be used to find the correlation between the contact forces and magnet displacement.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Estimating 2D Position from Magnetic Sensor Readings Using Artificial Neural Network

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Vladimir Sibinović,
  • Mirko Raković,
  • Milutin Nikolić,
  • Vladimir Mitić

摘要

This paper presents a systematic approach to data acquisition and analysis for estimating the 2D position of a permanent magnet using neural networks. In a step-by-step manner, the data were collected using a functional prototype, and the training was analyzed. Multiple neural network architectures were evaluated, and the best possible results were selected. This most effective network was exported as API and subjected to further real-world tests, and the results are presented. This research was conducted in an effort to create a sensor that can measure contact forces for robotic hands and grippers by measuring the change of the magnetic field of magnets suspended in a flexible substrate. The idea is that the methodology developed for determining the position of the magnet can be used to find the correlation between the contact forces and magnet displacement.