Microsoft AirSim on Unity Streamlines Autonomous Vehicle Training

Article by Limarc Ambalina | July 05, 2019

To provide developers with an easily accessible and safe way to train autonomous vehicle models, Microsoft launched AirSim on Unity. While autonomous vehicles have gained a lot of traction recently, research and development in the field can be arduous. Depending on where you are located, not only can it be difficult to test or train autonomous vehicle models on public roadways, but it may be completely prohibited. Due to safety concerns, driving autonomous vehicles on public roads is illegal in most parts of America and many other countries around the world.

However, high-quality AI training data and viable training areas are necessary to make numerous computer vision models function. Depending on the scope and scale of your project, Microsoft AirSim could be the ideal solution.


What is Microsoft AirSim?

How Microsoft AirSim on Unity Streamlines Autonomous Vehicle Training - Platform

Microsoft AirSim is an open-source robotics simulation platform. With the hope of providing another way to create training data and a space for training machine learning models, AirSim has numerous features and elements beneficial for computer vision development. The program was launched on Unity in late fall of 2018.


Machine Learning meets Industry-leading Game Development Platform

Unity is the video game development engine on which RollerCoaster Tycoon World, Pokemon Go, and numerous other augmented reality and simulation games were built. The Unity engine supports 2D, 3D, AR, and VR development. The platform has been highly acclaimed in the video game industry for its stunning graphical capabilities, making it an ideal virtual space for simulation and autonomous vehicle training.

Oneiros, a startup VR company based in Milan, created a photorealistic model of a Scandinavian house which shows the graphical power the Unity engine.


The Benefits of Working on Microsoft AirSim

Training autonomous vehicles or drones in the real world can be costly and often comes with risks which do not exist in a virtual training space.

  • Risk-free – Test your algorithm in a virtual world to work out any issues or bugs before testing it in the real world. Training your model in a virtual space means you don’t have to find or build training locations. As well, there is no risk of damaging hardware like drones, cameras, or vehicles during the training process.
  • Realism – Together with the Unity platform, rendered testing locations and environments are incredibly detailed and realistic on AirSim. Adjust shadows and reflections to create the ideal environments and train your algorithm to recognize minute visual details.
  • Powerful Tools & Compatibility – AirSim can interface with a variety of common robotics platforms and can work as a plugin to the Unreal Engine 4 editor from Epic Games. Unity boasts a large content library in their asset store, uses a C# development environment, and can run on both Windows and Linux operating systems.


How do you access Microsoft AirSim?

AirSim on Unity is still in early development and is currently listed as a beta release. However, it is fully accessible on Github with full installation and build instructions.

If simulation on AirSim isn’t an option for your project, Lionbridge provides AI training data and data annotation services for some of the world’s largest tech companies. Get in touch with Lionbridge AI to learn how our crowd of 500,000 expert annotators can help you get the data you need for your computer vision models.


Multilingual Data Annotation Services

Lionbridge provides professional data annotation services in over 300 languages.

Some of our most popular languages include:

  • Chinese data annotation
  • Italian data annotation
  • Dutch data annotation
  • Japanese data annotation
  • French data annotation
  • Portuguese data annotation
  • German data annotation
  • Spanish data annotation
Interested? Get high-quality data now
The Author
Limarc Ambalina

Limarc writes content for Lionbridge’s website as part of the marketing team. Born and raised in Canada, Limarc’s love of Japanese pop culture brought him to Japan in 2016 and living in Japan has been his dream come true. Apart from Lionbridge content, you can catch Limarc online writing about anime, video games, and other nerd culture.


    Sign up to our newsletter for fresh developments from the world of training data. Lionbridge brings you interviews with industry experts, dataset collections and more.