Safe Learning in Industrial Environments

Safe Learning in Industrial Environments

Safe learning in industrial environments focuses on developing algorithms that maximize the system’s performance and efficiency, while ensuring safety constraints are met and safe exploration is enabled.

Key aspects:

• Risk-aware learning and decision making
• Integration of safety constraints into AI and machine learning specifically in manufacturing or connected applications
• Real-world deployment and validation in industrial scenarios

Projects

• Safe learning for robot arms (manipulation of deformable objects)
• Multi-objective optimization in industrial environments (dynamic production scheduling and optimization)

Funding

• SNF - NCCR AUtomation
• Rieter Stiftungsprofessur
• Horizon Europe - DMaaST

Publications

• Autogeneration and optimization of pick-and-place trajectories in robotic systems: A data-driven approach M Wu, A Rupenyan, B Corves, Robotics and Computer-Integrated Manufacturing, 2025
• Efficient safe learning for controller tuning with experimental validation M Zagorowska, C König, H Yu, EC Balta, A Rupenyan, J Lygeros, Engineering Applications of Artificial Intelligence, 2025
• Safe time-varying optimization based on gaussian processes with spatio-temporal kernel J Li, M Zagorowska, G De Pasquale, A Rupenyan, J Lygeros The 38th Annual Conference on Neural Information Processing Systems (NeurIPS 2024)