Posts

Showing posts from January, 2021

AR Sandbox Part 13: Now What?

Image
  monitor and computer sit next to projector on its mounting slot With standard bad timing, the pandemic hit while discussions were being made with the middle school about displaying the sandbox in its lobby. Properly displaying this will have to wait, although I am certain good things can be done with it.  Until then, It will have to stay where it is. Thank you for reading! -Ryan Ellis (GSF)

AR Sandbox Part 12: It Works!

Image
  With the new hardware setup, the already installed code ran like a charm (although there were a few kinks) and the sandbox was now functional!    It is now fully reactive to the topography of the sand, and coolest of all: water simulation! You can hold your hand near the kinect, and water will be simulated falling from it, responding dynamically to the topography, and forming realistic bodies of water, flowing down mountains in handmade rivers into bodies of water.   Mission success! And for far less than the $15,000 that these things normally cost.  Now what to do with it?

AR Sandbox Part 11: Computer Rebuild

Image
 Don't let it's looks deceive you! This Dell Vostro case has much more under the hood than you would expect.   It contains the same components as the previous build with a few exceptions: New motherboard compatible with a more standard form factor New non-proprietary CPU cooler New NVIDIA GTX 970 GPU, more compatible with Linux and the software I am using (bottom)   Making this build and sourcing components was no easy task, due to their extreme age (at least as computer hardware goes)   But the end result performed beautifully. It showed itself to both resolve previous compatibility problems, improve reliability, and allow the PC much greater mobility and protection.

AR Sandbox Part 10: Hardware and Software Come Together

Image
Now that both the PC and the sandbox themselves were functioning, it was time to bring them together! As you can see, the kinect and software were synergizing well, and were able to generate a 3d model of the topography of the box. After more calibration, the projector was keyed in, and the system was able to make topographic projections onto the sand. But there was a problem. The frame rate was so incredibly poor, that you could not actually determine if the sand was adjusting to the right level. Water simulation was also proving to be impossible. After hours of command line performance troubleshooting outside in freezing temperatures, I decided that it was time to try a different brand of GPU. Around this time I also began to run into issues with he motherboard I was using. The computer would fail to turn on on occasion, which was likely due to a faulty motherboard, as adjusting capacitors often fixed it. Based on this, I elected to rebuild the computer again with a new GPU and

AR Sandbox Part 9: Computer changes

Image
  With the physical sandbox done, it was time to get back to the computer itself. I re-cased it in a flat plastic storage unit, and 3D printed an IO panel for it.    The power button was also on the outside of the "case". I had hoped this new case would improve the PC's isolation from the environment, and it was partly successful in this.    top view   A very nice close up  

AR Sandbox Part 8: Sandbox Build

Image
After finally getting all of the parts together, it was time to build the sandbox itself. All in all, this took over 10 hours, but yielded a very sturdy and functional result.           Base of sandbox projector/IR Scanner arm positioning the arm Attaching first set of casters   Bolting on the arm Finished sandbox!