Argonne Researchers to Develop Learning-Based Robots as Step Toward a Scientific Assistant

What if robots could not only perform experiments but also adapt and improve alongside human scientists?

LEMONT, Ill.--(BUSINESS WIRE)--Researchers at the U.S. Department of Energy’s (DOE) Argonne National Laboratory are developing a new way for robots to learn and adapt to many different hands-on laboratory tasks. The goal is to create robots that can work alongside scientists in real lab environments and adjust to changing conditions.

“Robots with fine motor skills already exist but using them safely and effectively in real laboratories is still very challenging,” said Nicola Ferrier, senior computer scientist. “Our approach starts by learning directly from expert scientists as they do their work.”

The RoSA: Robot Scientific Assistant for Accelerating Experimental Workflows project is part of DOE’s Genesis Mission, a bold national initiative to double America's research and development productivity within a decade. The key is harnessing artificial intelligence (AI), quantum computing and world-leading supercomputers.

As a first step, researchers will outfit fellow scientists with sensors and observe them as they prepare for and perform lab procedures. The recorded data will then be used to train computer models that allow robots to mimic expert actions and learn how tasks are performed correctly.

Ferrier brings experience in using computer vision systems to guide robots and machines. Her collaborator, computational scientist Arvind Ramanathan, has worked on self-driving laboratories and AI systems that can make complex decisions.

“Our main goal is to strengthen the basic robotics and computing tools needed so that large-scale, automated robotic systems can carry out experiments faster and more reliably,” Ferrier said.

Ramanathan said the techniques developed as part of the RoSA project will complement other research efforts such as Orchestrated Platform for Autonomous Laboratories (OPAL). That multi-lab project will create a network of autonomous laboratories that can learn and adapt, to accelerate breakthroughs across biology, biotechnology and energy science.

“In OPAL, dexterous robotics – which are well coordinated and nimble – are being planned for executing biological experiments,” he said. “By integrating AI-driven decision-making with advanced robotics, we aim to create systems that can accelerate discovery across a wide range of scientific disciplines.”

RoSA will also organize common lab tasks by how difficult and precise they are and map them to the most suitable type of robot. Fixed station robots have a stationary base and perform tasks within a defined workspace, whereas humanoid robots are mobile systems designed to resemble and move like the human body. Hybrid robots combine aspects of both. The project team will test robot performance in a virtual lab environment.

“Within the next year we hope to show a fivefold improvement in how efficiently these tasks can be completed,” Ferrier said. “In the long term, we envision robot scientific assistants that can work with existing laboratory equipment, making complex experiments both safer and more efficient. RoSA is a key step toward that future.”

The work is funded by DOE’s Office of Science, Advanced Scientific Computing Research program.