Alisa Rupenyan

Alisa Rupenyan

she/her

Professor, industrial AI for autonomous manufacturing systems

Sitemap

A list of all the posts and pages found on the site. For you robots out there, there is an XML version available for digesting as well.

Pages

Posts

End of 2025 update

less than 1 minute read

Published:

2025 was an amazing year for our group. We tripled in size, in terms of group members, and in projects. We started work on new topics, such as building energy management via learning control (collaboration with the Swiss industry), and dynamic manufacturing scheduling approaches (Horizon EU project). We have further developed the Learning Robotics Lab with several new systems: Winding robot, AGV, and a couple of collaborative robots for manufacturing applications, also focused on deformable object manipulation. Several publications came out (non-exaustive list):…

Robotics Lab growing

less than 1 minute read

Published:

Our robotics lab is growing. Here are some recent impressions.

Bridge Discovery Grant Accepted!

less than 1 minute read

Published:

Our collaborative grant focused on robotic 3D-printing has been accepted and jointly funded by the Swiss National Science Foundation (SNSF) and the Swiss Innovation Agency (Innosuisse). It is a 4-year project for 2 MCHF, and a collaboration with Efe C. Balta at inspire AG and with John Lygeros at the Automatic Control Lab at ETH Zurich.

End of the year 2024

less than 1 minute read

Published:

What an year that was. The group grew with two more people (although they will join in a few weeks), we published several cool publications in top venues (such as NeurIPS and Engineering Applications of AI for AI, and in IEEE TCST and IEEE TASE for control and learning). I actively joined the second phase of NCCR Automation, with one PhD project, one postdoc project, and one industry collaboration.

publications

Safe risk-averse bayesian optimization for controller tuning

Published in IEEE Robotics and Automation Letters, presented on IEEE ICRA 2023, 2023

RAGoOSe is a novel data-driven approach that combines safe learning with risk-averse Bayesian optimization to safely tune controllers in high-precision systems with variable noise, demonstrating improved performance over traditional methods in both synthetic and real semiconductor manufacturing applications.

Recommended citation: Christopher König, Miks Ozols, Anastasia Makarova, Efe C Balta, Andreas Krause, Alisa Rupenyan (2023). "Safe risk-averse bayesian optimization for controller tuning." IEEE Robotics and Automation Letters. 8(8208 - 8215).
Download Paper | Download Slides

Robotics and Manufacturing Application

Published in The Impact of Automatic Control Research on Industrial Innovation: Enabling a Sustainable Future, John Wiley & Sons, 2024, 2024

A book chapter, focused on advanced control methods, suitable for industrial applications in robotics and precision motion systems.

Recommended citation: Alisa Rupenyan, Efe C. Balta (2024). "Robotics and Manufacturing Application." The Impact of Automatic Control Research on Industrial Innovation: Enabling a Sustainable Future, John Wiley & Sons, 2024.
Download Paper

Time-varying safe data-driven optimization

Published in The Thirty-Eighth Annual Conference on Neural Information Processing Systems (NeurIPS 2024), 2024

We propose a method for optimization with unknown and time-varying optimization objective and constraints, based on Bayesian optimization.

Recommended citation: Jialin Li, Marta Zagorowska, Giulia De Pasquale, Alisa Rupenyan, John Lygeros, NeurIPS, 2024
Download Paper | Download Slides

Efficient safe learning for controller- and process parameter tuning

Published in Engineering Applications of Artificial Intelligence, volume 143, 1 March 2025, 109894, 2025

We propose a method for safe and fast data-driven optimization by formulating a series of optimization problems instead of a grid search.

Recommended citation: M Zagorowska, C König, H Yu, EC Balta, A Rupenyan, J Lygeros Engineering Applications of Artificial Intelligence, 2025
Download Paper

In-situ controller autotuning by Bayesian optimization for closed-loop feedback control of laser powder bed fusion process

Published in Additive Manufacturing, Volume 99, 2025

We employ Bayesian optimization as an automated, sample-efficient method for tuning in-layer controllers in laser powder bed fusion additive manufacturing.

Recommended citation: Barış Kavas, Efe C. Balta, Michael R. Tucker, Raamadaas Krishnadas, Alisa Rupenyan, John Lygeros, Markus Bambach, Additive Manufacturing, Volume 99, 2025
Download Paper

In-situ controller autotuning by Bayesian optimization for closed-loop feedback control of laser powder bed fusion process

Published in Additive Manufacturing, Elsevier, volume 99, 5 February 2025, 104641, 2025

We experimentally apply Bayesian Optimization in LPBF to tune an in-layer PI controller to modulate laser power, assessing its performance on wedge geometries prone to overheating.

Recommended citation: Barış Kavas, Efe C Balta, Michael R Tucker, Raamadaas Krishnadas, Alisa Rupenyan, John Lygeros, Markus Bambach, Additive Manufacturing, volume 99, 2025
Download Paper

Adaptive Bayesian Optimization for High-Precision Motion Systems

Published in IEEE Transactions on Automation Science and Engineering, Volume 22, presented on IEEE IROS 2025, 2025

We propose a real-time purely data-driven, model-free approach for adaptive control by online tuning low-level controller parameters using safe Bayesian optimization.

Recommended citation: Christopher König, Raamadaas Krishnadas, Efe C. Balta, Alisa Rupenyan, IEEE Transactions on Automation Science and Engineering, Volume 22, Pages 15627–15637, 2025
Download Paper

Bayesian Optimization for Automatic Tuning of Torque-Level Nonlinear Model Predictive Control

Published in 2026 European Control Conference (ECC), accepted, 2026

We present an auto-tuning framework for torque-based Nonlinear MPC using high-dimensional Bayesian Optimization and a digital twin, achieving significant improvements in trajectory tracking on a UR10e robot arm.

Recommended citation: Gabriele Fadini, Deepak Ingole, Tong Duy Son, Alisa Rupenyan, European Control Conference (ECC), 2026
Download Paper

Guided Multi-Fidelity Bayesian Optimization for Data-driven Controller Tuning with Digital Twins

Published in IEEE Robotics and Automation Letters, vol. 11, no. 5, 2026

We propose a guided multi-fidelity Bayesian optimization framework integrating corrected digital twin simulations with real-world measurements for data-efficient controller tuning in closed-loop robotic systems.

Recommended citation: M. Nobar, J. Keller, A. Forino, J. Lygeros, A. Rupenyan, IEEE Robotics and Automation Letters, vol. 11, no. 5, pp. 5294–5301, May 2026
Download Paper

research

Autonomy - performance

Published:

Methods for increasing the performance of manufacturing systems under unknown constraints (read more…)

Industrial robotics

Published:

Our research in industrial robotics enables robot-based systems to autonomously and precisely perform manufacturing-related tasks

research_group

Our Research Group

Published:

Welcome to the Industrial AI Research Group. Our team focuses on developing solutions for smart manufacturing and industrial systems, in particular industrial autonomous robotics, data-driven optimization for manufacturing, and integration of AI-based methods in various aspects of manufacturing.

talks

Keynote, Entrepreneurship in Control

Published:

I co-organized and chaired the tutorial session on Entrepreneurship in Control of the European Control Conference 2025 in Thessaloniki and gave a keynote about presenting research with entrepreneurship in mind - drawing examples from our research in control and optimization of 3D printing. Here is more info about this tutorial. The website contains many useful resourses - in particular, how to find applications and how to extend the scope of your research. My slides are available upon request.

Conference talk, Adaptive run-to-run optimization for precision motion systems

Published:

In manufacturing and industrial environments, environment or equipment is subject to change. By using a method based on Bayesian Optimization and safe exploration, our method optimizes continuously desired parameters based on the prescribed system performance. We include input and output constraints, which are satisfied throughout the optimization procedure. Therefore, the method is suitable for use in practical systems where safety or operational constraints are of concern. We include contextual information, using task parameters, which makes the optimization flexible. The paper can be found here on arxiv and as a journal article in IEEE Transactions on Automation, Science, and Engineering here.

Keynote, Flexible automation for robot manufacturing tasks

Published:

During the workshop Embodied AI and Edge Computing for Intelligent Robots at IROS I presented several our advances in controlling manufacturing robots to reach autonomy and adapt to various manufacturing scenarios. One of the important technologies to reach that is metalearning, on which we work actively. Using interactive optimization (Bayesian optimization, reinforcement learning), and adapting models or controllers in the cloud enables even better adaptation capabilities. I also highlighted an upcoming publication (Multifidelity Guided BO with DIgital Twins), which also demonstrates how edge and cloud computing increase autonomy and adaptability for robotic and manufacturing systems and components.

Innovation Booster Robotics as an Enabler in the Swiss Robotics Innovation Ecosystem

Published:

I had the honor to present the Innovation Booster Robotics in front of the robotics community in Switzerland on this year’s Swiss Robotics Day, and to talk about its achievements and about its excellent team. I am grateful to work with the booster for a number of years.

During the Swiss Robotics Day, our group participated with the ZHAW pavilion and showed a demonstration of our robotic winding system, working seamlessly for 10 hours with path planning computed by a RL algorithm.

teaching

Artificial Intelligence 1

Undergraduate course, ZHAW, Computer Science, 2024

This is an introductory course on the basics of Artificial Intelligence, with a strong focus on applications. It covers topics from deep learning and from planning and search.

MLDM

Undergraduate course, ZHAW, Computer Science, 2024

This is an introductory course in Machine Learning. The material is organized around the topics described here.