Counting 2,409 Big Data & Machine Learning Frameworks, Toolsets, and Examples...
Suggestion? Feedback? Tweet @stkim1

Last Commit
Feb. 22, 2018
Jul. 4, 2017


Build Status Simulation Status

We choose to go to the moon in this decade and do the other things,
not because they are easy, but because they are hard.
-- John F. Kennedy, 1962

Welcome to the Apollo GitHub.

Apollo is an open autonomous driving platform. It is a high performance flexible architecture which supports fully autonomous driving capabilities. For business contact, please visit

Apollo Team now proudly presents to you the latest version 2.0.


We strongly recommend building Apollo in our pre-specified Docker environment. See the following instructions on how to set up the docker environment and build from source. The steps are:

  • Run a machine that runs linux (tested on Ubuntu 16.04 with and without an nVidia GPU)
  • Create a docker environment
  • Build Apollo from source
  • Bootstrap start Apollo
  • Download the demonstration loop and run it
  • Start a browser session and see the Dreamview user interface

More instructions are below

The docker environment can be set by the commands below.

Running a docker container from a linux system is done with this sequence. At the end, the script will get you into the container

bash docker/scripts/
# logout and login back in the system to make sure to run docker command without sudo
docker ps  # to verify docker works without sudo
# in case you forgot to logout and login back, do so, remove ~/.docker/config.json
# and check again with `docker ps`
bash docker/scripts/
bash docker/scripts/

To build from source

First check and make sure you are in development docker container before you proceed. Now you will need to build from the source. If you want to run the entire system, make sure you have an nVidia GPU and that you have installed the Linux nVidia drivers.

# To get a list of build commands
# To make sure you start clean
./ clean
# This will build the full system and requires that you have an nVidia GPU with nVidia drivers loaded
bash build

If you do not have an nVidia GPU, the system will run but with the CUDA-based perception and other modules. You must specify either dbg for debug mode or opt for optimized code

./ build_no_perception dbg

If you make modifications to the Dreamview frontend, then you must run ./ build_fe before you run the full build.

Run Apollo

Follow the steps below to launch Apollo. Note that you must build the system first before you run it. Note that the will actually succeed but the user interface will not come up if you skip the build step.

Start Apollo

Running Apollo will start the ROS core and then startup a web user interface called Dreamview, this is handled by the bootstrap script, so from within the docker container, you should run:

# start module monitor
bash scripts/

Access Dreamview

Access Dreamview by opening your favorite browser, e.g. Chrome, go to http://localhost:8888 and you should see this screen However, there will be nothing running in the system.

Access Dreamview

Replay demo rosbag

To see if the system works, use the demo 'bag' which feeds the system.

# get rosbag note that the command download is required
bash ./docs/demo_guide/ download

# You can now replay this demo "bag" in a loop with the '-l' flag
rosbag play -l ./docs/demo_guide/demo_2.0.bag

Dreamview should show a running vehicle now. (The following image might be different due to changes in frontend.)

Dreamview with Trajectory


Apollo documents can be found under the docs repository.

Ask Questions

You are welcome to submit questions and bug reports as Github Issues.

Copyright and License

Apollo is provided under the Apache-2.0 license.


Please refer the Disclaimer of Apollo in Apollo official website.

Latest Releases
 Dec. 29 2017
 Sep. 19 2017
 Jul. 4 2017