- Electrical muscle stimulation synced to foot movements recreates the feel of flat, uphill and downhill surfaces

- Compact wearable setup replaces large treadmills, with potential uses in VR gaming, education and training

An overview of the EMSWalker system, a virtual reality walking platform in which a seated user alternately raises and lowers the front of each foot to move a virtual avatar, while electrical muscle stimulation is applied to the leg muscles activated during real walking to simulate the sensation of traversing flat, uphill and downhill terrain. [Provided by GIST]
An overview of the EMSWalker system, a virtual reality walking platform in which a seated user alternately raises and lowers the front of each foot to move a virtual avatar, while electrical muscle stimulation is applied to the leg muscles activated during real walking to simulate the sensation of traversing flat, uphill and downhill terrain. [Provided by GIST]

Researchers have developed a virtual reality technology that lets users feel as though they are walking uphill or downhill — while seated, without a large treadmill, simply by moving their feet.

The Gwangju Institute of Science and Technology (GIST) announced Wednesday that a research team led by Professor Kim Seung-jun of its AI department, working jointly with researchers at the Massachusetts Institute of Technology's Computer Science and Artificial Intelligence Laboratory (CSAIL), has developed a virtual walking system called EMSWalker.

Conventional VR can render a virtual environment that looks realistic, but a gap remains between what appears on screen and what the body actually feels. Walking on flat ground, uphill and downhill each engages different muscles at different moments, making it difficult to replicate a genuine sense of walking simply by moving one's feet while seated.

The research team reasoned that stimulating the muscles used in walking — in sync with foot movements, even without actual walking — could bridge that gap.

EMSWalker links a "foot-tapping" motion to movement in the virtual environment: the user sits and alternately raises and lowers the front of each foot, and the virtual avatar steps forward when the foot lifts and plants when it lowers. Walking speed in the virtual space changes with the pace of the foot movements.

The key element is electrical muscle stimulation (EMS). In time with each foot movement, electrical impulses are delivered to the major muscles involved in walking — the front of the thigh, the shin and the calf. The timing and intensity of the stimulation are calibrated to match the different patterns of muscle use on flat, uphill and downhill terrain.

When the team had participants identify terrain type through muscle stimulation alone, without showing them any walking animation on screen, participants correctly identified uphill terrain 78 percent of the time, flat terrain 72 percent of the time and downhill terrain 58 percent of the time — all well above the 33 percent chance rate of randomly selecting one of three terrain types.

Applying muscle stimulation also heightened users' sense of "embodiment" — the feeling that the virtual avatar is their own body. When stimulation was varied to match flat, uphill and downhill terrain in a full VR environment, users reported a stronger sense of "presence," or the feeling of actually being inside the virtual space, compared with other stimulation methods. No significant differences were observed in VR motion sickness or physical and mental discomfort.

The GIST-MIT international research team. From left: Kim Seung-jun, professor in GIST's AI department; Eom Ju-won and Kang Seong-jun, doctoral candidates; Kim Bo-cheon, a doctoral candidate in the computer science department at the University of Chicago; Kang Yu-min and Kim Gwang-bin, doctoral candidates in GIST's AI department; Shin Se-mu, a researcher at GIST's Artificial Intelligence Research Institute; Joseph DelPreto, a researcher at MIT's Computer Science and Artificial Intelligence Laboratory; and Professor Daniela Rus. [Provided by GIST]
The GIST-MIT international research team. From left: Kim Seung-jun, professor in GIST's AI department; Eom Ju-won and Kang Seong-jun, doctoral candidates; Kim Bo-cheon, a doctoral candidate in the computer science department at the University of Chicago; Kang Yu-min and Kim Gwang-bin, doctoral candidates in GIST's AI department; Shin Se-mu, a researcher at GIST's Artificial Intelligence Research Institute; Joseph DelPreto, a researcher at MIT's Computer Science and Artificial Intelligence Laboratory; and Professor Daniela Rus. [Provided by GIST]

The research team highlights that the system can reproduce the sensation of walking using only compact sensors and a wearable electrical stimulation device, with no large treadmill or incline apparatus required. They say it could serve as a locomotion interface for home extended-reality (XR) setups where space is limited, as well as for VR games, immersive content, and education and training applications.

The team plans to optimize the system for individual users' stride length, gait habits and sensitivity to electrical stimulation, and to extend it beyond uphill and downhill surfaces to simulate walking on varied terrain such as sand and mud.

"We have confirmed the possibility of varying the physical sensation of walking to match changes in terrain seen in a virtual environment," Kim said. "We plan to expand the technology to express a wider range of slopes and surface characteristics, enabling a richer experience of virtual movement."

The findings are set to be presented at IEEE ISMAR 2026, an international conference on augmented and mixed reality, to be held in Bari, Italy, from Oct. 5-9.


nbgkoo@heraldcorp.com