Beta Testing Kinect Task 4

What was the configuration of the computer(s) on which you verified the software? (manufacturer, model, processor type, processor speed, system memory, operating system version)
Hewlett-Packard (HP 620)
Windows 7 Home Premium- Service Pack 1
Pentium® Dual- Core CPU T4500 st 2.30GHz
4.00 GB (2.90 GB Useable)
32- bit Operating System

Which software development languages did you use? (LabVIEW? C/C++? Java?)
LabVIEW

What problems or difficulties did you encounter?
Well at first we were insure what to do and how to make our own gestures. However with the help of some much needed clarification through examples we were able to figure it out. Well commented examples and explanations of the Kinect Vis really were key to helping us understand everything.

What questions did you have during the process?
We were constantly questioned ourselves as to whether it would work. Luckily in the end it did.

What would have made this task easier?
The "Kinect Gesture Examples" under file releases made it easier. The 'Kinect Gesture Processing' VI served as a lovely example that provided a good start.. We also used the 'Joints to Angle' VI heavily in our programming. We feel that is a very helpful and universal VI that many teams could use. So we would suggest adding as a premade VI within the WPILib Kinect API.

Any specific suggestions on improving the documentation?
For the Kinect joints you are given a list of twenty options. For some of them we are unsure what they are referring to. For example, one of them is 'Head'. What do you do with your head? We don't know. It would be really handy if there were diagrams of what each joint option referred to. We as a team are considering doing something like that and (once we made the document) would share it as a reference guide.

Were the provided examples and documentation covering the WPILib Kinect API straightforward? Were they sufficient to allow you to begin experimenting with modifying the gestures and/or creating your own?
Normally when you hover over a VI with help on it gives you some information. Since this is still in beta it didn't. We mentioned this early on and Mr. Greg McKaskle provided helpful descriptions. As long as those are added in, then the documentation will be much improved.

Did you try implementing your own gestures using the WPILib Kinect API? If so, did you find you preferred them to the canned gestures? Would you be willing to provide a video of a user utilizing these gestures, and a brief description of the gestures and how they mapped to robot functions in addition to the code?
We did make our own gestures. We currently are doing some very simple stuff and are planning to continue testing with perhaps more complex gestures. The teleop portion of our code is attached. Basically each arm controls a side of the robot. By making a 90 degree angle with your elbow, either pointing up or down, you make the robot drive forwards or backwards, respectively.

Did you try making modifications to the example gestures? Did you feel these modifications were an improvement? If so, what did you change and how did it improve the existing control?
Well right now we find that the system is a bit too sensitive. We are going to add in a subVI that will throw out the extremes. For example, anything under 20 degrees would be ignored and treated as zero degrees. We are hoping this will make the system less jerky.

Is there anything else you want to tell us related to this task?
When we were testing the premade gestures that are accessed like Joystick buttons we found out that your arms had to be out in order to access the gestures. While we understand that this works as a safety check, it was a little awkward to have your arms out while moving one of your legs back and forth.
This is more about the Kinect in general, but we will add it here. While we do find that it is a very cool feature, we see it more as a novelty. We still feel safer and more comfortable driving our robot with joysticks. It really goes back to the fact that controlling a robot with human movements tends to be a bit more unstable. While the robot can be controlled with changes in a joint in degrees, humans do not move their joints in 10-degree increments. While the field of gesture controlled robotics is fascinating, within this scenario the human body is not exact enough to provide the kind of control you might want within a FIRST competition. We do, however, enjoy showing it to people as an example of the possibilities of robotics.