GroundedReach: Enabling Body-grounded Haptic Experience in Virtual Reality with an Elbow Wearable Haptic Device

Reach out and press your palm flat against a wall. Your hand is stopped by the wall from being stretched, and you can feel the resistance force of your palm, your wrist, your elbow, your shoulder, and, furthermore, down your spine. Potentially, to render such an experience in Virtual Reality (VR), we may need to recreate all those forces on the user's body. This may require the user to wear a complex exoskeleton that fits her body dimensions, and may be cumbersome to put on or take off.

Implementing such hardware is expensive and may be impractical for consumer use. A new research paper out of the VR Haptics groups at the University of Birmingham and the Southern University of Science and Technology shows that it may be enough to render only part of the haptic — for example, only on one of the arm joints — to support the generation of a mental model of pushing the object. The brain, tasked with building a single unified model that explains the world around the user, can fill in for the missing haptic rendering.

Such top-down decisions of the brain may generate mistakes, as can be demonstrated in many optical illusions, where we can see geometry or motion that does not exist — yet in haptics they can be used to generate compact and simpler devices, as shown in this work.

The group tested participants and applied haptic feedback to selected joints of the arm: the elbow, the shoulder, an empty palm, or one holding a VR controller, and different combinations of these, while participants were asked to push virtual boxes of different simulated weights.

The results showed that applying resistance to at least one joint of the arm and holding a VR controller in the palm created the experience that the virtual box had mass that resisted the user's push. Adding resistance on additional joints of the arm did not raise the realism of the experience in a measurable way. Realism only dropped when there was no arm resistance at all. The brain, it turns out, is quite willing to fill in the rest of the arm once the elbow gives it a strong enough anchor.

The Grounding Problem

Current handheld controllers used in VR have long been used to render a short buzz or a rumble to notify the user that their hand has touched a virtual object. What they can't do is stop your hand from moving, penetrating the object. Push against a virtual wall with a standard controller, and your arm sails right through it. The illusion breaks precisely at the moment it matters most: contact with something solid.

Devices that can stop your arm — full-body exoskeletons, room-scale robotic props, floor-based shape displays — do it by building a mechanical bridge from your limb all the way down to the ground or the walls of the lab. They work, but they're heavy, expensive, and take real effort to put on, fit, and calibrate. That's a lot of infrastructure for something as simple as a doorframe or a control panel that refuses to move.

The team noticed that a large share of everyday VR interactions — reaching for a switch, pressing a button, bracing against a wall — happens while the user is standing or sitting more or less in place, extending an arm outward. In that narrow but common scenario, you might not need to ground the whole body. Maybe grounding one joint is enough.

GroundedReach is a single, compact unit worn like a brace around the elbow, easily put on the arm or taken off with Velcro strips. The device combines different mechanisms together to generate the haptic feedback.

A magnetic powder brake provides continuously adjustable resistance — enough to render the difference between nudging a light drawer and hauling on a stiff one, as well as rendering more complex changing resistance profiles. An additional servo-controlled directional ratchet handles the other end of the spectrum: it locks and arrests the joint rotation when the virtual object is meant to be immovable. A small load cell tucked into the design keeps tabs on how hard the user is actually pressing, so the firmware can release the brake the instant someone eases off, without waiting on a round-trip command from the VR application. The design aims to make the device disappear — the resistance should feel like it belongs to the object, not to the hardware on your arm.

Does It Actually Help?

The team ran two follow-up studies. The first looked at a simple task: push a virtual box forward until it hits a wall at a given distance, as quickly and accurately as possible, comparing visual-only feedback, controller vibration, and using the elbow device.

The effect was clear. With the elbow device, participants located the wall using noticeably less movement, in a fraction of the time. Participants' comments get at something the numbers alone don't capture: the device changes the strategy people use, not just their speed. Instead of visually hunting for a boundary and correcting as they go, participants simply reached until the arm stopped for them — freeing up visual attention for whatever else the virtual scene demanded.

For the second study, the team built a demo dungeon game with three rooms, each leaning on a different one of the device's capabilities: eight doors where only one opens (rigid stopping), a corroded valve wheel that must be cranked to raise water (rotational resistance), and a set of hammers of different weights used to strike an ore (impact and weight rendering).

Every room, with haptics switched on, scored significantly higher on both perceived realism and enjoyment than the same room experienced with visuals alone.

The haptics doing something more useful than just feeling nice — it helped them understand how to interact with an object in the first place.

Several participants described the haptics doing something more useful than just feeling nice: it aided comprehension of object interaction strategies. Without resistance cues, one participant noted, it was genuinely hard to tell which of two visually similar doors was meant to open. With elbow haptic rendering, the difference was immediate and physical.

Such a device still has limitations. It is a purely passive system: the brake and ratchet can resist or arrest motion, but they can't push back or shove your arm the way an active exoskeleton could. And because it only constrains the elbow, a determined user can still rotate around the shoulder and drift through a virtual wall if they really try.

None of that erases the central finding, though: for the reaching, pushing, and bracing motions that make up so much of everyday VR interaction, one well-placed joint carries most of the perceptual weight. The rest, the brain is happy to supply on its own.

Small Device, Wide Doorway

The small elbow device is easy to put on. It sidesteps the custom-fitting and setup burden that has kept full exoskeletons largely in the lab. The authors see it as a complement to existing controllers rather than a replacement — controllers keep doing fine-grained buttons and vibration, while the elbow unit supplies the one thing they've never been able to offer: a wall that actually stops your hand.

Participants, unprompted, started imagining their own uses for it — rowing games, punches landing in a fighting game, the pull of an elastic band in a home workout. It's a reminder that once a joint like the elbow is given a voice, people are quick to find things for it to say.