Salzburg University of Applied Sciences and ABB's Machine Automation Division (B&R) have filed a joint patent for an AI-based method to improve energy efficiency in industrial drive systems, including robots, machine tools, and automated production lines. The work, conducted at the Josef Ressel Center for Intelligent and Secure Industrial Automation (JRZ ISIA), targets the challenge of energy losses that are difficult to model precisely using conventional control methods.
Key takeaways
- Reinforcement learning (RL) agents learn directly from real system behavior, adapting motion control strategies without requiring a complete system model.
- A novel mathematical formulation of the learning strategy accelerates training and reduces data requirements, making RL viable for industrial deployment.
- The approach aims to make motion sequences—such as positioning, acceleration, and cyclic movements—substantially more energy-efficient while reflecting real operating conditions.
- The research builds on work initiated in 2020 under the EU Interreg project KI-Net and has been developed since 2022 with industry partners including B&R and COPA-DATA.
This collaboration highlights a practical pathway for integrating advanced AI research into industrial automation, potentially enabling more energy-efficient operations in cyber-physical systems. The patent filing underscores the value of academic-industry partnerships in delivering tangible innovations for manufacturing.
Source: Robotics & Automation News (roboticsandautomationnews.com) · Published 2026-06-03 · “ABB and Salzburg researchers patent AI system to cut energy use in industrial robots”