Amazon has recently open-sourced the DeepRacer device software, the software used to run AWS DeepRacer. The 1/18th scale autonomous vehicle and the DeepRacer events allow developers to create machine learning models and race them in a cloud based 3D racing simulator.
David Smith, senior solutions architect for AWS DeepRacer, explains why Amazon decided to open source the project:
The AWS DeepRacer virtual and in-person leagues have been a hit, but now developers want to go beyond league racing with their car. Because the AWS DeepRacer is an Ubuntu-based computer on wheels powered by the Robot Operatingk System (ROS) we are able to open source the code, making it straightforward for a developer with basic Linux coding skills to prototype new and interesting uses for their car.
AWS DeepRacer was introduced in November 2018 and runs on AI models using reinforcement learning techniques as explained in the AWS Open Source blog:
Reinforcement learning (RL) has become one of the most popular machine learning techniques for training robots in simulation. RL enables models to learn complex behaviors without labeled training data and allows the models to make short-term decisions while optimizing for longer-term goals.
The open source AWS DeepRacer software will be available on GitHub, with Amazon providing 15 Robot Operating System (ROS 2) packages and documentation to help developers build their applications. Among the projects, sample applications provided by Amazon (including Follow the Leader, Off Road, and Mapping) and projects submitted by contributors. Tomasz Ptak, senior software engineer at Duco and AWS DeepRacer league finalist, recommends some of the projects from the AWS DeepRacer Community:
I’ve seen how this project evolved over time. I especially recommend the sample projects from our Community members, Martin Paradesi Jochem Lugtenburg and Chris Miller.
Amazon started the AWS DeepRacer League, in person and virtual races where developers get experience with machine learning in a fun but competitive environment. Marcia Villalba, senior developer advocate at AWS, explains how Amazon recently reorganized the competitions to encourage new ML users to get involved with the platform:
Over the past year, the AWS DeepRacer League’s races have gone completely virtual and participants have competed for different kinds of prizes. However, the competition has become dominated by experts and newcomers haven’t had much of a chance to win. The 2021 season introduces new skill-based Open and Pro racing divisions, where racers of all skill levels have five times more opportunities to win rewards than in previous seasons.
InfoQ previously covered AWS DeepRacer and DeepLens, a video camera for developers, with a presentation by Jeremy Edberg, CEO and Founder of MinOps, at QCon.ai.
To promote the open source projects, Amazon temporarily reduced the prices of the AWS DeepRacer, AWS DeepRacer Evo, the second generation car equipped with stereo cameras and LiDAR sensor, and the AWS DeepRacer sensor kit.