Main Cluster: ITEA 3
Second Cluster: PENTA-EURIPIDES
High-tech cyber-physical systems (CPSs) play increasingly important roles in our society. They are ubiquitous, and companies, organizations and societies depend on their correct functioning. CPSs need to have high up-times, be user-friendly, and economically to use. CPS suppliers must assure that their systems reliably deliver optimal quality in customers’ environments, without bothering their customers with complex system optimisation tasks that require highly skilled staff. Systems need to be optimally tuned before delivery and at installation and re-adjusted during use, which can easily require many hours/days and this total time increases rapidly due to growing project diversity and complexity. To address this major problem, it is ASIMOV’s vision that CPSs must be increasingly autonomous and self-optimising, which leads to the following central question:
How to build complex high-tech systems that select their optimal settings autonomously within minimal time and with minimal external expertise? To answer this question, the ASIMOV project will develop innovative technologies to create self-optimising CPSs by combining AI and Digital Twinning. The consortium, consisting of large industrial parties, SME’s with strong AI-expertise, and leading universities and research institutes, will deliver the following innovations:
• creating digital twins of systems to simulate realistic system behaviour;
• training an Optimisation-AI based on the digital twin to find optimal system settings;
• verifying the validity of the digital twin for training the AI;
• using the trained AI to perform the tuning and calibration tasks on actual machine configurations.
This will lead to AI-based software that autonomously performs system optimisation tasks during manufacturing, installation, and system usage. Proof of concepts will be provided in three different industrial system domains (electron microscopes, automated driving, process control) for which optimisation is crucial for system performance.