Skip to content

chankh/donkey

 
 

Repository files navigation

donkeycar: a python self driving library

Build Status CodeCov PyPI version Py versions

Donkeycar is minimalist and modular self driving library for Python. It is developed for hobbiests and students with a focus on allowing fast experimentation and easy community contributions.

Quick Links

donkeycar

Use Donkey if you want to:

  • Make an RC car drive its self.
  • Compete in self driving races like DIY Robocars
  • Experiment with autopilots, mapping computer vision and neural networks.
  • Log sensor data. (images, user inputs, sensor readings)
  • Drive your car via a web or game controler.
  • Leverage community contributed driving data.
  • Use existing harsupport supportdware CAD designs for upgrades.

Getting driving.

After building a Donkey2 you can turn on your car and go to http://localhost:8887 to drive.

Modify your cars behavior.

The donkey car is controlled by running a sequence of events

#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)

See home page, docs or join the Slack channel to learn more.

About

self driving car

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Python 57.3%
  • JavaScript 35.7%
  • HTML 4.6%
  • Shell 1.2%
  • CSS 1.2%