As the ELO-X Project is approaching its final phase, a new video summarizing the project’s motivation, goals, and impact has been released by the Consortium.
Archive
New learning-based MPC project SLIMPEC kicks off
Implementing model-based control systems in industrial applications presents two significant challenges: developing accurate yet robust mathematical models and calibrating controller parameters to ensure optimal performance, not only in ideal conditions but throughout the product’s lifecycle. These processes are often time-intensive, technically demanding, and costly.
To overcome these challenges, the new project SLIMPEC will create advanced software tools that leverage machine learning and control system theory. These tools aim to:
- Construct physics-informed predictive models that seamlessly integrate white-box process knowledge with data-driven learning and adaptive techniques, and
- Establish (semi-)automated calibration methods to deliver near-optimal parameter tuning offline while enabling real-time adjustments to adapt to evolving operating conditions.
The project, kicked-off on November 1st, 2023, and expected to complete by October 31st, 2026, is carried out by ODYS and SUPSI within the Eureka Eurostars programme of Horizon Europe.
ODYS QP solver and MPC software presented at IFAC workshop
Dr. Gionata Cimini gave a talk on “Embedded optimization and MPC for next-gen controls in industrial production” at the Virtual Seminar IFAC TC on Optimal Control titled “Trends on software tools for optimal and predictive control”. The event was held on May 18, 2022.
This has been a great opportunity to present ODYS QP solver and MPC products, and discuss our recent advancements in learning-based predictive control. We showed results of benchmarks carried out on an embedded board, comparing our software with other well-known tools for optimization.
We are proud to say that ODYS solver was found to be the fastest, most reliable, and most light on memory of all tested tools in the considered dataset for embedded applications.
Check out the video of the presentation below and have a look at the slides.