As more industrial and autonomous robots come online for an increasingly diverse array of applications, ensuring adequate safety protections exist for the humans who share their workspaces is a growing concern. Luis Duffaut, an assistant professor in the Department of Electrical and Biomedical Engineering, was recently awarded a prestigious to continue his research to develop advanced control systems using a data-driven approach to establish safety-control laws while continuing to optimize and guarantee performance.

According to the , over a half million industrial robots were installed worldwide in 2021, with nearly 35,000 new installations in the United States alone. The world’s current total stock of industrial robots is estimated at 3.5 million units.

Capable of lifting heavy objects, working in hazardous environments, and handling repetitive tasks with great speed and precision, modern robots have demonstrated the potential to reduce the frequency of ergonomic injury to workers. However, having these large and powerful machines laboring side-by-side with humans can create new, unfamiliar hazards. A 2023 research study published in the identified 41 workplace fatalities as being robot-related in the 25 years from 1992 to 2017.

“Expanding autonomous industrial applications require safe control systems that allow tasks to be performed efficiently while taking advantage of an abundance of captured data, enabled by progress in sensor miniaturization and fast processing,” said Duffaut, “My long-term research vision is to relieve this need by advancing the field of autonomy and robotics from both data-driven and safety perspectives while addressing pressing societal challenges, producing new knowledge, and accelerating public acceptance.”

Assistant Professor Luis Duffaut demonstrates a robot's Lidar System
Assistant Professor Luis Duffaut demonstrates a four-legged robot's Lidar (Light Detection and Ranging) System which is a remote sensing method that uses light in the form of a pulsed laser to measure distances.

His CAREER research project calls for a 5-year plan to establish the fundamental theories for safety and data-driven systems within a universal and systematic framework. The resulting autonomous technologies blend core disciplines in physics and mathematics—which happen to be the respective programs in which Duffaut received his bachelor’s and master’s degrees—with advanced electrical and control engineering.

“Here in electrical engineering, we don't build robots. We build the algorithms that make them work autonomously. We work with things that we cannot observe— with things that are abstract, and challenging to conceptualize,” said Duffaut. “That is what we teach our students— how to think abstractly so that they can create the algorithms to ensure a robot is efficient, accurate, and safe.”

Like an artist who has replaced his brushes with mathematics, Duffaut describes the complex algorithms he and his students develop as possessing an inherent beauty. “Engineering gives a purpose for the mathematical theory,” he added. “I look for things that are beautiful and at the same time, solve a problem. Being beautiful means that they may be efficient, they may be easy to understand, they may be easy to apply and implement on an algorithm in the shape of a program.”

The advanced autonomous technologies that Duffaut is pursuing would utilize new frameworks for data-driven systems to explore innovative methodologies that allow for both safety and performance guarantees.

It is a process that often begins unceremoniously via a process he calls decomposition, where an existing or theoretical system is deconstructed, or simplified, to its fundamental components and then carefully reassembled in such a way that allows researchers to extract all the information from the system to reveal all its hidden properties.

Assistant Professor Luis Duffaut in the AIRLab
Assistant Professor Luis Duffaut shares a device developed by an undergraduate student that uses compressed air to replicate the motions of a satellite in space.

The resulting methodology and estimators, often feature a complex balancing act between in signal processing, data discrimination, real-time machine learning, and the use of redundancy to arrive at a system that does not compromise safety for the sake of efficiency, nor the opposite—where a system is bogged down and slowed by inefficient controls.

This is the second CAREER Award this year for faculty in the Department of Electrical and Biomedical Engineering (EBE). Assistant Professor Amber Doiron was also named a CAREER awardee for her innovative research in developing methods and materials for advanced wound dressings that better combat serious infections. The EBE Department now boasts four CAREER Award winners with Associate Professor Mads Almassalkhi being recognized in 2021 and Associate Professor Hamid Ossareh receiving his CAREER Award last year.

“We are thrilled to see our faculty recognized for their exceptional innovation in developing solutions to complex real-world challenges,” said Department Chair and Professor . “These awards not only celebrate their pioneering research but also their ongoing commitment to inspire and prepare the next diverse generation of STEM leaders.”

Educational Impacts:

With a data-driven approach to further advance and optimize autonomous control systems, Duffaut’s project is not limited to developing technological advancements but also embraces a broader ambition of improving educational and societal impacts. He hopes the ground-breaking research opportunities will attract and inspire new generations of engineering students who share his motivation “to address the challenges related to a world of human-machine interactions, such that these interactions benefit society in a responsible and ethical manner.”

By advancing the field of autonomy and robotics from a uniquely data-driven and safety-focused perspective, Duffaut hopes the research project will also benefit the industry by accelerating public acceptance of the developing technologies.

“Our goal is to capture the imagination of students, the new generations of engineers,” he said, “For our automation students, it does not matter if it is a robot with wheels or arms or is a tiny aerial drone. We view everything under the language of systems.”

With the research and educational components in place, Duffaut sees potential for ¶¶Ňő̽̽ to be the autonomy hub of Vermont. By cultivating mutually beneficial partnerships with local industry leaders like , , , and , the University can further support the state’s thriving tech sector that recently gained a strong endorsement by the federal government when designated as .

Luis Duffaut tests a robot in the AIRLab
Demonstrating the remarkable agility of modern robots, a four-legged robotic friend of Assistant Professor Luis Duffaut leaps into the air.

As a project with the potential to enhance the department's reputation for autonomy as an area of focus, Duffaut and his colleagues see ample opportunities to deepen interest in STEM education in a state where young minds are already fascinated with the promise of robotic technology. This past February, the College of Engineering and Mathematical Sciences helped to host the 2024 Vermont First Tech Challenge (FTC) as 19 teams of middle and high school students from across the state participated in the annual robotics competition.

Reflecting on the honor of being named an NSF CAREER Award winner, Duffaut is quick to recognize and extend his appreciation to ¶¶Ňő̽̽’s Office of the Vice President of Research (OVPR), for their assistance and to his department’s chair, Marilyn J. Cipolla, saying, “She has been fundamental to the development of this project and quintessential in her support of faculty as they evolve as researchers.”