• Competition website
• Competition guidelines
• Linux and Windows Binaries
• Python API
  • airsimneurips @PyPI
  • airsimneurips API doc

This competition marks the advent of a new release process for AirSim, in which we have separated out the AirSim plugin from environment content. Instead of having a series of individual executables for each environment, we have compacted all of the relevant AirSim content and API into a single binary (AirSimExe).
Unreal environments containing race courses are released as separate downloadable content (DLC) packages, in the form of .pak files, which can be loaded and unloaded into the main binary as needed.

• Navigate to the releases page.
• Download the latest Windows or Linux AirSim.zip.
• Download your desired environments (pakfiles) - Building99.pak / SoccerField.pak / ZhangJiaJie.pak.
• Move the environment pakfile into AirSim/AirSimExe/Content/Paks.
• The contents in the environment pakfile will now be linked to your AirSim binary!

• Linux

  • Open a terminal window, cd to AirSim/ directory, and enter the following command:

     ./AirSimExe.sh -windowed
    

• Windows

  • Navigate to the AirSim/ directory, and double-click run.bat

To control your drone and get information from the environment, you will need the airsimneurips API, which is accessible via Python.

• To install the Python API for the Neurips competition, please use:

   pip install airsimneurips
  

  Corollary: Do not do a pip install airsim, as we will have a few custom APIs specific to this competition.
• Resources
  • airsimneurips API doc
  • AirSim upstream API and examples
    (Note that this is not used in the competition, however is a good learning resource)

There are two ways to swap between levels, either via AirSIm API or by the UI menu.

• API

  • We have added a new API simLoadLevel(level_name="MainMenu") to change Unreal environments on the fly.
    Possible values for level_name are : "Soccer_Field_Easy", "Soccer_Field_Medium", "ZhangJiaJie_Medium", "Building99_Hard". Here's a quick snippet to iterate throught them all. Before trying this, please ensure you've downloaded the corresponding pak files from our releases page.

   import airsimneurips as airsim
   client = airsim.MultirotorClient()
   client.confirmConnection()
   client.simLoadLevel('Soccer_Field_Easy')	
   client.simLoadLevel('Soccer_Field_Medium')	
   client.simLoadLevel('ZhangJiaJie_Medium')
   client.simLoadLevel('Building99_Hard')

• UI Menu

  • Press F10 to toggle the level menu
  • Click your desired level. (Note: the UI lists all the pakfiles in the AirSim/AirSimExe/Content/Paks directory. Ensure you downloaded the pakfile, if you are not able to see a particular environment)

• Plan and move on minimum jerk trajectory using gate ground truth poses:
  • Generate an AirSim settings.json file

  $ cd baselines;
  $ python generate_settings_file.py

  • Start the AirSim Neurips binary, as explained above
  • Run the code!
    See all the baseline arguments here

  $ python baseline_racer.py \
  	--viz_traj \
  	--plot_transform \
  	--planning_baseline_type all_gates_at_once \
  	--planning_and_control_api moveOnSpline \
  	--level_name ZhangJiaJie_Medium 

• Soccer Field: A simple outdoors environment with few obstacles, and an easy to follow course.
• ZhangJiaJie: A mountainous landscape based on a national park in the Hunan province of China.
• Building99: A tight race course designed inside one of Microsoft's very own buildings.

Please open a Github Issue on this repository (not AirSim) for any technical questions w.r.t. the Neurips competition.