We wanted to build a roomba with local vision processing for real time targeting

What it does

Shoot nerf darts at all who cross it's path

How we built it

We connected the Kinect to the Raspberry Pi, and started getting them to work together. This involved setting up and installing various libraries and firmwares, and some source compilation to get some of the older more outdated libraries to work. Then we started getting vision processing to work on the Pi, and getting a skeleton from the Kinect. This involved further more library searching, compilation, and testing. In the meanwhile, we also had to develop a mount and firing mechanism for the Nerf gun. This required multiple 3d printed parts to be designed and printed, as well as some modification to the body of the Nerf gun. We also had to find ways for C and Python to communicate data from one to the other in a manner they both understood.

Challenges we ran into

3d Printing connecting Kinect to Raspberry Pi Installing drivers for Kinect on Raspberry Pi Installing skeleton tracking software on the raspberry Pi Getting Skeleton information into Python Automating Nerf gun firing Designing mounts Soldering wires Kinect is often deprecated

Accomplishments that we're proud of

We managed to make a lot of discontinuous libraries communicate and send each other information, and use said information

What we learned

The Kinect is an older piece of hardware, and has outdated libraries to match. Getting old libraries to work takes a lot of leg work.

What's next for CleanShot

Share this project: