Skip to content

smellslikeml/LitterBug_donkey

 
 

Repository files navigation

LitterBug: Autonomous Trash Rover

Py versions

The LitterBug is a trash pick up rover built off the Donkeycar platform. It's goal is to help pick up trash in beaches, parks, etc as well as provide a fun and relevant context for which to learn about self-driving vehicles, AI, and robotics.

Read these first!

donkeycar

Build a LitterBug to:

  • Create a version of a real life WALL-E!
  • Experiment with autopilots, mapping, computer vision, and neural networks.
  • Train your rover to pick up trash anywhere (..and have fun while saving the planet)
  • Log sensor data. (images, user inputs, sensor readings)

Start Training!

After building a Litterbug, press the select button on the ps3 controller and get rolling.

Create your own parts.

You can make the LitterBug interact with the environment in new ways by adding a new physical part. For example, this project added a scoop part using an extra servo. Since the LitterBug is built off the Donkeycar platform, you can use their structure to create new parts.

#Define a vehicle to take and record pictures 10 times per second.

from donkeycar import Vehicle
from donkeycar.parts.camera import PiCamera
from donkeycar.parts.datastore import Tub


V = Vehicle()

#add a camera part
cam = PiCamera()
V.add(cam, outputs=['image'], threaded=True)

#add tub part to record images
tub = Tub(path='~/mycar/get_started',
          inputs=['image'],
          types=['image_array'])
V.add(tub, inputs=['image'])

#start the drive loop at 10 Hz
V.start(rate_hz=10)

You can also control your new part by mapping an unused button on your PS3 controller. Modifying the donkeycar/parts/controller.py file:

#AVAILABLE BUTTONS = ['tr', 'tl2', 'tr2', 'mode', 'thumbl', 'thumbr']
#AVAILABLE AXES = ['y', 'z', 'rx']

#In the update() method of the JoystickController class add:

if button == '<AVAILABLE BUTTON HERE>' and button_state == 1:
    #do something!

See LitterBug write-up, donkeycar resources, or join the Donkeycar Slack channel to learn more and share your build.

Packages

No packages published

Languages

  • Python 55.6%
  • JavaScript 36.6%
  • HTML 4.9%
  • Shell 1.4%
  • CSS 1.2%
  • Dockerfile 0.3%