[apalrd]

I talked with Bryan (BJC) and Nick (dsm33) directly, and Jim Zondag via email about this today. As we are a beta test team, we only have to wait until Monday to see our Kinect.

We talked about possibilities of the Kinect on the robot and DS end.

We all agree that using the Kinect on a driver station is a very very bad idea, as joysticks are much more direct in use, and the driver can command specific operations easily.

Some of us think that it could be useful on the robot end, however:
-It is not a light sensor (weight-wise). It has two cameras, an IR laser, four microphones, and a motorized base (yikes!)
-It cannot interface directly to the cRio, requiring at least a single-board computer to interface to. For weight reasons, we obviously want a processor that dosen't need a large heat sink, and the lowest-power processors aren't likely to run windows 7. Luckily, there are linux drivers for the Kinect, and its embedded nature means we can run without a local interface (no GUI or graphics processor requirement).
-This brings our total requirements to:
*cRio (in kit)
*Kinect (in kit)
*Single-board computer with embedded linux programming skills (???) - OR - larger single board computer capable of windows 7 which requires much more power and cooling and .NET programming skills (???)
*A challenge which can't be solved any other way to make this all reasonable (this really scares me)


I think its much more reasonable to assume that the Kinect is provided because Microsoft is trying to promote it, and probably donated it, and that the use for it will be negligible. Plus, how do they expect it to work on the field with multiple robots and Kinects (as the Kinect would likely be confused by other IR patterns from other Kinects, especially 6 operating at once).