Grant Funding Yields Promising Results for ECE Researcher

There has been significant progress on machine-learning research being conducted at the University of Wyoming in the year since it was funded.

University of Wyoming Assistant Professor Domen Novak and his research team are using a $447,889 National Science Foundation grant, awarded in 2017 and running through 2020, to examine how machine learning can be used to recognize human workload and stress from physiological responses. UW Psychology Professor Sean McCrea and his team also are involved in the project, which is entitled, “A Kinder, Gentler Technology: Enhancing Human-Machine Symbiosis Using Adaptive, Personalized Affect-Aware Systems.”

In the project, human participants are asked to perform different activities at different difficulty levels. Sensors affixed to participants’ provide real-time physiological measurements such as heart rate, and machine learning techniques are then used to recognize the participants’ level of workload. Based on this information, the computer can adjust task difficulty or modify content as necessary in order to ensure a more positive experience for the participant.

Novak says the early research results indicate usefulness in several applications, but it’s necessary to extract more data and determine the best way to use it.

“If we can incorporate pre-existing user characteristics – say someone is already neurotic or someone else has a relaxed personality to begin with – how do you recognize how the machine adapts to that?” Novak says.

This research is part of the broader area of affective computing, which is the study and development of systems and devices that can recognize, interpret, process and simulate human affects.

“What we’re seeing is that personality characteristics do matter, and results do indicate that we can better tailor tasks to the user than previous methods,” Novak says. “It’s paving the path for widespread use of physiological-input machine learning.”

Previously, Novak was awarded separate grant funding to improve driver attention spans with the goal of reducing motor vehicle fatalities. In collaboration with Associate Professor Mohamed Ahmed of the Department of Civil and Architectural Engineering, Novak conducted a study where test subjects “drove” in UW’s Driving Simulator while outfitted with physiological sensors with various distractions, such as texting.

Novak’s research at UW includes human-robot interaction with an emphasis in rehabilitation robotics. He is interested in the information that robots can obtain about human performance, intentions and emotional states via sensors such as heart rate, electromyography, electroencephalography and eye-tracking. His area combines robotics, biomedical signal processing, sensor fusion and virtual reality.