AI-based control of a skiving post-processing machine using imitation learning
Overview
- Project Term
- 2024 - 2025
- Leadership
- Raphaela Erbel
Prof. Dr. Philipp Brune
Günter Ihle (Firma Rigdon GmbH (opens in a new window))
- Status
- in planning
Project description
The cooperation project with the company Rigdon GmbH (opens in a new window) aims to optimize the tire retreading process using machine learning. The challenge is to automate the complex and labour-intensive reworking step within the skiving process in tire retreading. The aim of this project is to develop a control algorithm for a robot based on imitation learning. Imitation learning is an approach in the field of machine learning in which a model learns to perform a task by imitating human experts. Imitation learning uses direct observations of successful strategies. This is often contrasted with classic reinforcement learning, in which learning takes place through trial and error and the application of a reward function. The method of imitation learning is particularly useful in situations where defining a clear reward function is difficult or where exploring the solution space can be dangerous or costly - both of which apply to this application scenario.
Given the fact that 571,000 tons of used tires are generated annually in Germany alone, this project not only improves the efficiency of the retreading process and thus the competitiveness of retreated tires compared to new tires, but also offers great potential environmental relief potential.
In summary, this project not only represents a technological innovation and improvement, but also addresses key social challenges such as the shortage of skilled workers and environmental protection.