Sloane: Dog Finder

We noticed there are lost dog fliers for Sloane, a three legged dog. So we built a Dog Finder robot to find lost dogs

Comment

Inspiration

Sloane is a three legged dog and without her family. Her family has been looking for her for the last few days, posting fliers around UCLA. We noticed these fliers right outside the Hackathon, and decided to relief the stress of the owners by help finding their dog using a robot.

What it does

Named after the lost dog, Sloane, our robot, uses A* (pathfinding algorithm) and google cloud services auto ML to find specific lost dogs. Sloane has a stereo camera and uses openCV as parameters for the ML api.

How we built it

3d printing, a star, google cloud services auto ml, raspberry, ...

Challenges we ran into

  • Missing sensors/parts: there were no gyroscope sensors (or manometers), so we couldn't use SLAM for movement correction; there were also no encoder available.
  • WiFi and local networks: connecting to the rpi and the rpi to the internet for cloud services were slow and difficult making testing and prototyping last longer.
  • 2 Motors and wheels: resource constraints caused us to make a 2 wheel drive with a ball bearing roller since there were no more wheels left.

Accomplishments that we're proud of

Sloane can recognize the dogs in general. But Sloane can also differentiate between the dog its looking for and any dog it finds. Sloane is mostly 3d-printed and works wirelessly with a battery.

What we learned

Dog finding is difficult, especially training a network to differentiate between to similar dogs or dog looking objects.

What's next for Sloane: Dog Finder

Finding more dogs!

Share this project:

Updates