You Can't Walk in a Straight Line—And That's Great for VR

Redirected walking can trick people into believing they’re exploring huge areas when they’ve been walking in circles in a room.
VRWalkingstory
Then One/WIRED

Humans, quite simply, suck at walking in straight lines: Give a man a blindfold, and he will walk around in circles. In virtual reality, though, this failing becomes very convenient. People are so bad at knowing where they are in space that subtle visual cues can trick them into believing they’re exploring a huge area when they’ve never left the room—a process called redirected walking.

With VR technology getting ever better, people could one day explore immersive virtual spaces—like buildings or even whole cities—on foot in head-mounted displays. But it's not very immersive if you end up smacking into a real-life wall. “This problem of how you move around when in VR is one of the big unsolved problems of the VR community,” says Evan Suma, a computer scientist at the University of Southern California.

That’s where redirected walking comes in. People don’t really notice, it turns out, if you increase or decrease the virtual distance they had to walk by 26 percent or 14 percent, respectively. Or shift the virtual room so they see their path as straight when their real path is curved. Or turn the room up to 49 percent more or 20 percent less than the rotation of their heads.

Research in Action June 2015Stephanie D. Kleinman

How are we so oblivious to these mismatches? It might seem especially odd when you consider the motion sickness associated with VR. Just a tiny fraction of second mismatch between movement seen and movement felt can cause nausea. “Timing is critical,” says Suma. Your eyes, muscles, joints, and vestibular system—that’s the fluid in your inner ear whose sloshing creates the perception of motion—all work together.

But as long as you see a movement and sense a movement, the magnitude or direction of that movement doesn’t have to be precise. “Your non-visual systems are sloppy,” says Eric Hodgson, a psychologist at Miami University. In the real world, visual cues correct any sloppiness in sensing movements. (Just think of all the times you’ve walked through your living room—yet you still bang your shins trying to walk through it in the dark, says Hodgson.) But in virtual reality, software controls all the visual cues, tricking people's suggestible limbs into walking a path that makes sense.

Your body is still good at reacting to those visual cues quickly, though, which is why hardware solutions to moving in VR will be difficult to implement. Omnidirectional treadmills lag by virtue of having to move physical parts rather than just refresh pixels on screen. They’re also expensive. Another solution is good old joysticks—but that takes away part of VR’s immersiveness. So redirected walking might be the most effective way forward.

At left, a diagram of the virtual space in blue and the actual space in red, to scale. At right, the path through the virtual room. Razzaque et al / EUROGRAPHICS 2001

Razzaque et al / EUROGRAPPHICS 2001

The first paper to show redirected walking in VR came from Mary Whitton’s lab at the University of North Carolina at Chapel Hill. Folk wisdom has long held that people lost in the desert or in the woods end up walking in circles, but virtual reality was too jankity to test it out for a long time. In the 2001 study, Whitton and her graduate students put eleven people in a virtual fire drill.

Before exiting the virtual hallway, they had to zigzag to reach and push four buttons—each turn was an opportunity to redirect. The virtual space was a long hallway, but the actual room was less than half the size, and participants ended up doubling back on their route rather than turning at an angle. “I still remain surprised how well it works,” says Whitton.

A door is rotated 90 degrees in virtual space. Suma et al 2010

Suma et al 2010

More sophisticated redirection techniques are appearing, too. Suma has taken advantage of a phenomenon called change blindness: If we don’t see an object move, we don’t really notice its position change. For example, he rearranges the virtual world behind a participant's back by rotating the location of a door by 90 degrees. (It may not surprise fans of Portal that the video game inspired Suma’s research into spatial manipulation.) He tested 77 participants who each went through a dozen such rotated doors—and only one person noticed anything funky going on. By using this technique of rotating doors, he’s able to create a virtual building of several thousand by several thousand feet in a 16 by 16 room.

Redirected walking works well in the lab, but it’s still a ways off from the living room. For one, you still need a large empty space (read, no furniture) for walking around. The other constraint is money: Redirected walking requires very good position tracking of a person over a relatively big space; these systems exist, though they’re still too expensive for the average customer. But hey, that’s what everyone was saying about VR headsets just a few years ago, too.