Kinect Technologies for PCs, can track you through walls

Intel’s gesture control tech will be built into PCs from 2014

Ever since Microsoft’s Kinect came out, it has been wondering when the technology would get built into PCs. Yes, there is Kinect for Windows, but it’s a peripheral — about having advanced motion detection capabilities in the webcam, as a bridge to exciting future user interfaces.

Well, today, such technology is on its way, but not from Microsoft. No, it’s Intel that the PC manufacturers are talking to, and it’s not Kinect that’s the base: it’s Intel’s perceptual computing technology.

According to Paul Tapp, senior product manager in Intel’s perceptual computing division, manufacturers have “committed to doing it” in 2014 – “it” being the integration of an Intel-designed motion-detection system into their machines. And in the meanwhile, peripherals maker Creative put its $210 Senz3D, the first retail device to use the technology.

Creative’s Senz3D camera is up for pre-order. It’s the first peripheral to use Intel’s perceptual computing tech, which will be built into computers from next year. >>

Contributed by By David Meyer Gigaom.com

MIT’s ‘Kinect of the future’ can track you through walls

Researchers from MIT have unveiled a new form of motion tracking that uses a three-point system to follow a person’s position, even through a totally opaque wall. Though the word “Kinect” has been thrown around quite liberally for the sake of accessibility, this is strictly a positional tracker — that means that it won’t be interpreting sign language or reading lips any time soon. Rather than being a control mechanism, this device is purely for keeping tabs on users as they move both within and between rooms. At present the tracker is set up directionally, so it can only see through the single wall at which it is pointed, but the obvious end goal is an omnidirectional tracker that could follow a user through the whole house, upstairs and down.

The system works using three radio antennas spaced about a meter apart to bounce signals off a person’s body. Even through the researchers’ office wall, it can follow people with an accuracy of up to 10 centimeters (four inches), better than WiFi localization can currently provide. Though the device is exploded and sitting as component parts at present, one grad student working on the project said they expect to be able to condense it down to a final unit no larger than Microsoft’s Kinect sensor.

Beyond the loss of Kinect-like image and silhouette tracking, the MIT system can also only track a single person at a time. A second moving object within the system’s field of view will cause confusion and make the system useless — though that problem is, of course, to be addressed soon. It also has trouble with stationary objects, but they already have a first pass on an algorithm to get around this by recognizing the motion of a person breathing.

Applications for the technology, assuming its kinks and limitations are addressed, are numerous. There are the obvious gaming applications, perhaps blurring the line between real and virtual locations as players stalk through real hallways full of video-game enemies. All Oculus Rift fantasies aside though, there are plenty of more substantive reasons to be excited about the ability to keep track of people without their need to carry a transmitter. Rather than installing motion trackers in every corner of the home, a single tracker near the center might be able to intelligently turn the lights on and off as you move from room to room.

Architects and advertising researchers would love to know how people move through a particular space, where they spend their time, and what places they tend to avoid. The health care industry could keep better track of people in need of supervision, receiving an alert if, say, a person with dementia begins to wander away.

Of course, there are also the more troubling possible uses. WiFi localization currently requires users to hold a tracking device, while more versatile options like holographic localization are slow and low fidelity. MIT is now bringing a high degree of accuracy and usability together with the versatility that comes with being able to track people who have never consented to be tracked. If the signal could be made strong enough, it could render prison break-outs virtually impossible, or let law enforcement quickly check the number and position of people in a hostage situation.

Human and civil rights activists might have something to say about such applications, however. That’s really the downfall of a catch-all people-tracker for use outside of private homes: I can’t imagine a world in which its use would remain legal for long. People are leery enough about ad agencies tracking their online activities — how might people react to the idea of a company monetizing their walking path through the local mall? The Kinect has already got certain people up in arms over just the possibility of always-on functionality, and that would only have mattered when the user was standing directly in front of their television.

The team has a patent pending for the technology, but the concept seems like it would be easy enough to adapt with slight changes. It’s still in its infancy, but finding a person through a wall by picking up on their breathing is about as strong a proof of concept as they could ever have hoped for.

Contributed by Graham Templeton Extremetech.com

Related
 Xbox One Kinect can understand two people speaking at once, PS4 counters with camera bundle

Related post: 
LiFi, instead of WiFi: Chinese scientists achieve Internet access through lightbulbs