Focus on Didactic Factories: INDUSTRY4.0LAB's experiment

WHAT ABOUT THE EXPERIMENT? 

Industry4.0lab, a Didactic Factory at Politecnico di Milano (POLIMI) in Lombardy, aims to provide hands-on training for MSc Engineering students, addressing gaps in traditional curricula. MADE s.c.a.r.l. is actively engaged in collaboration with this experiment, bringing forth its expertise as one of the eight Italian competence centers and a recognized Digital Innovation Hub. 

​

The facility houses a manufacturing line for assembling small electronic devices and features robotic assets for material handling and end-of-life tasks. Human-Robot Interaction (HRI) is a growing trend in manufacturing, with collaborative robots (cobots) gaining attention for enhancing human-based operations. However, cobots require manual setup adjustments, posing challenges in highly flexible environments like electronic board de-manufacturing. 

To address this, an experimental setup integrating camera vision and machine learning (ML) algorithms is proposed. ML will cluster boards based on physical dimensions and compatibility with handling equipment, streamlining hardware setup decisions. Additionally, task allocation between humans and robots will be optimized based on component size, enhancing collaboration efficiency. 

Synthetic data generation will augment testing diversity, overcoming limitations posed by limited real-world board availability. This fusion of HRI, cobot technology, and ML aims to boost DF productivity and enrich student education, offering solutions for efficient, flexible electronic board manufacturing and advancing Industry 5.0. 

The experiment targets improved efficiency in highly flexible tasks, focusing on handling-oriented clustering and data-driven cobot and end effector selection. Expected outcomes include reduced board handling errors, decreased setup time, minimized variance from worker expertise, and enhanced multidisciplinary skill development for students.  

 

WHAT IS EXPECTED?  

The scenario envisions algorithms offering real-time instructions for selecting and mounting grippers on chosen cobots. Equipment from POLIMI and MADE facilities will be utilized initially in a virtual factory setting. The ML infrastructure's successful implementation will streamline operator tasks, preventing time wastage and minimizing stress from cobot handling failures.