CS Ph.D. Student Kyungyeon Lee Studies Haptics to Reshape How People Learn Physical Skills

From learning to play an instrument to performing a surgical procedure, mastering physical skills often depends on repetition and correcting mistakes over time. Researchers in human-computer interaction are examining whether technology can reshape that process by physically nudging users earlier in motor execution, rather than relying solely on post-error correction.

At the University of Maryland, Kyungyeon Lee, a Ph.D. student in the Department of Computer Science, is contributing to that effort through research to be presented on April 12 at ACM CHI 2026 in Barcelona, Spain. The paper, which received a Best Paper Honorable Mention Award and ranked among the top 5% of 6,740 submissions, explores how hand exoskeleton systems can support motor sequence learning through preemptive feedback.

“Having this work accepted at ACM CHI with an Honorable Mention is very meaningful,” Lee said. “Our work takes a different strategy from the mainstream that skill learning relies solely on guidance, so it was especially rewarding to see this approach recognized.”

Advised by Assistant Professor of Computer Science and Immersive Media Design Jun Nishida, Lee studies human-computer interaction, focusing on haptics, motor skill learning, and personalization. Her work centers on designing haptic systems that physically interact with users' bodies to augment their motor skills and learning efficacy, and adapt to individual learning patterns, particularly in tasks that require precision and coordination.

In the paper, Lee developed a hand exoskeleton called FIXical I/O (fix + physical + I/O), which explores how real-time error sensing can be paired with physical intervention. The system uses electromagnetic actuators and magnetic field sensors to detect and physically prevent users from making errors before they occur, with the entire sequence happening within 150 milliseconds.

In experimental studies, Lee found that preemptive feedback influenced how users approached learning, with participants showing increased confidence, greater awareness of potential errors, and improved learning performance compared with traditional feedback methods.

The study also raises questions about how training systems could be designed in settings where errors carry significant consequences. 

“I believe the implications of this work extend beyond the lab,” she said. “In high-stakes training environments, such as for pilots, surgeons, patients during physical therapy, or air-traffic controllers, errors can be costly or dangerous, and systems that intervene before a mistake is made could change how people are trained for these roles.”

As the field of human-computer interaction continues to evolve, Lee said researchers are increasingly focusing on systems that integrate more closely with the human body and adapt to individual needs.

“I design and evaluate haptic systems that support more adaptive and personalized physical skill acquisition,” Lee said. “I tackle this by combining personalization techniques with haptic devices to optimize the learning experience.” Lee previously published related research at ACM CHI 2025.

Looking ahead, Lee said her future work will continue to examine how physical skills are learned. 

“Learning new skills is inherently challenging but exciting,” she said. “I would like to establish new ways to make this process not only effective but also more empowering by enhancing users’ sense of agency and competence through personalization, and I hope that haptic technologies can augment our physical capabilities, embodied knowledge, and the quality of our lives, as AI continues to advance."

See more work on haptics, exoskeletons and even muscle stimulation at the Embodied Dynamics Laboratory.

—Story by Samuel Malede Zewdu, CS Communications