MIT is using millimeter wave technology to make VR headsets wireless

Shawn Knight

Posts: 15,296   +192
Staff member

First-generation virtual reality platforms from major players like HTC and Oculus VR have finally delivered on the promise of consumer-grade virtual reality albeit with some significant limitations.

As they stand today, the HTC Vive and the Oculus Rift must be tethered to a reasonably fast PC to drive the experience. Fortunately, there’s a small army of researchers, engineers and developers tirelessly working to both reduce the sheer amount of processing power required to drive flagship-class VR platforms and perhaps more importantly, cut the cord.

HTC just the other day announced an add-on device for the Vive called the TPCAST, a platform-specific third-party accessory that’s available on a very limited basis.

Researchers from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL), meanwhile, have been working on a solution of their own that’ll transform any tethered headset into a wireless solution.

Dubbed “MoVR,” the prototype system utilizes high-frequency millimeter waves, a much-ballyhooed technology with several potential applications. Its major shortcoming, however, is the fact that they don’t play nice with physical obstructions like walls. Even in an open room, you’d need perfect line-of-sight between a transmitter and receiver to support the sustained data rates necessary for virtual reality. Something as trivial as moving your hand in front of an equipped headset could disrupt the flow of data.

To overcome this downside, the team at MIT developed MoVR to act as a programmable mirror capable of detecting the direction of incoming millimeter waves and reconfiguring itself to reflect the waves toward the receiver on a headset. The team says its solution can learn the correct signal direction to within two degrees, allowing it to correctly configure its angles.

The current iteration consists of two directional antennas, each of which being less than half the size of a credit card, that utilize phased arrays to focus signals into narrow beams that can be “electronically steered.”

PhD candidate Omid Abari, who co-wrote a paper on the topic, says future revisions could be as small as a smartphone, thus allowing users to place several devices in a single room and enable local VR multiplayer.

Permalink to story.

 
Cool! I think... this could work I suppose with an antenna that sticks up above your head. unfortunately for what they are VR headsets are still out of reach for a lot of people in terms of price. Making them wireless is just another cost.
 
Sounds like they took tech that has been in radar for decades, and applied it to data transmission. The question is whether they can get the headset portion able to do similar beam forming and steering.
 
Sounds like they took tech that has been in radar for decades, and applied it to data transmission. The question is whether they can get the headset portion able to do similar beam forming and steering.
Why would it need to? The headset doesn't send any data back to the computer as far as I'm aware. Motion detection is all done externally, at least on the Vive. I can't remember but maybe the Oculus sends motion data back via USB cable...
 
Why would it need to? The headset doesn't send any data back to the computer as far as I'm aware. Motion detection is all done externally, at least on the Vive. I can't remember but maybe the Oculus sends motion data back via USB cable...
A lot of headsets are trying to get all motion-sensing done internally. Motion capture cameras only add to the cost and complexity. Once that happens - and it will - it will need a two-way signal. With this in an MIT lab right now, that means it might be ready for market right around the same time that internal motion sensing is standard on most VR headsets, and sits like Techspot starts taking points off in their reviews of headsets that still use an external system.
 
I see... although I would think the bandwidth needed to send motion data is considerably less than what is required to receive live video. So I imagine sending that motion data back could be done over a more common RF setup.
 
I see... although I would think the bandwidth needed to send motion data is considerably less than what is required to receive live video. So I imagine sending that motion data back could be done over a more common RF setup.
It isn't so much a question of bandwidth as just the act of simultaneously sending and receiving. You either need two antennas, one each for the transmitter and receiver, or a very well designed system and antenna that can handle pulling double duty.

It is MIT, so I have little doubt that if it can be pulled off with present knowledge, then they will be the ones to do it. But it will still be a challenge to get it small enough to rest comfortable on a headset.
 
Back