Challenging the traditional approach to human-robot interaction

Taylor Higgins, assistant professor of mechanical engineering at the FAMU-FSU College of Engineering, has co-authored an article that challenges conventional thinking about human-robot interaction.

The article, published in Science Robotics, argues that successful robotics development must focus more on understanding human behavior rather than solely on robot functionality. The research, by an all-female team, highlights fundamental differences in how humans and robots perceive their environment, make decisions and execute tasks.

“The key point from this paper is that for robots to improve their performance, they need to adapt smartly, sensing that humans are sometimes unpredictable and will change their actions accordingly,” Higgins said. “The idea is that both humans and robots can adjust their behaviors together over time.”

Collaborative origins

The project began in 2022 at the University of Texas at Austin, when Higgins, then a postdoctoral researcher, met Keya Ghonasgi, now an assistant professor at Rice University, who was completing her doctorate. Their collaboration expanded to include Meghan Huber, assistant professor at the University of Massachusetts, and Marcia O’Malley, professor at Rice University.

“I’m particularly proud of the results from this all-female author team,” Higgins said. “We wrote the paper, yes, but ended up supporting one another in the day-to-day wins and losses of the academic pipeline, which was an extremely enjoyable experience.”

Innovative research applications

Higgins leads several projects exploring human-robot interaction, including an unusual study of motor learning through unicycle riding.

“As whimsical as it sounds, unicycling is an ideal platform for studying motor learning because it requires three-dimensional balance as well as forward propulsion,” Higgins said. “By studying this unique skill, I’m diving deep into the motion that mirrors the art of walking.”

Her research also examines human intent prediction in environmental contexts, studying how humans anticipate and execute actions like sitting down when they spot a chair.

Advancing rehabilitation technology

The research has significant implications for rehabilitation robotics, particularly in devices like the EksoNR lower-limb powered exoskeleton by Ekso Bionics, which assists in gait training after neuromuscular injuries.

“Early controllers for devices like these are aimed to guide the user through normative gait patterns,” Higgins said. “However, scientists have found that the user needs to be actively engaged in attempting to initiate the movement themselves, otherwise they generally don’t realize any rehabilitation goals. So, we hope in the future by understanding current limitations we can improve these devices.”

This review article marks the beginning of what the researchers hope will be ongoing collaboration.

“Human-robot interaction is no longer a brand-new field so we challenge other researchers to take the next step in developing better human-robot interactions in the future,” Higgins said.