PrimeSense may be busy making its next-gen sensor smaller and not more powerful, but Microsoft isn't looking to stuff Kinect into mobile devices. The next generation Kinect is going to be an integral part of Microsoft's next gaming console, (widely expected to be released at E3 in June), and there's a rumor that it will come with some beefed-up specs, which we've got for you in a nice handy table. Will it make your robot better, smarter, and faster? Well, maybe.
We should reiterate that these specifications are rumored. A leak website got them from I know not where, but people with more experience than I have deemed them apparently legitimate. Here's the rundown on the new sensor (currently known as "Durango"), from VGLeaks:
The biggest disappointment here for robotics is likely the unchanged depth resolution. With a minimum resolvable depth of just 0.4 meter, robots that are too small to use Kinect to help them execute grasping tasks with arms and grippers close to their sensors aren't going to get much help from the Kinect 2.
However, we're definitely liking the increased resolution of the depth data, which should enable resolving smaller objects and features from longer distances, better separation of objects occluding one another, and better curve and edge recognition. According to the rumors we're basing all of this on, "at 3.5 m it can resolve objects two to three times smaller than the current sensor." Also, the USB 3.0 interface should make it possible for users to access all of these data, which was an issue with USB 2.0.
There are also rumors of improved skeletal tracking, which can track six people (up from two), can identify occluded joints, identify some joint rotations, and pick out thumbs and fingers well enough to be able to tell an open hand from a closed one.
Again, these are rumors, but they seem pretty legit. We'll be on the ground at E3 in June to get you all the details as the announcement happens.