Amazon releases DeepRacer software in open source

In November 2018, Amazon launched AWS DeepRacer, a car about the size of a shoebox that runs on AI models trained in a virtual environment with reinforcement learning techniques. DeepRacer has expanded since then, with new miniature race cars and a women’s league. And starting today, Amazon is making the DeepRacer device software available in open source.

The pandemic has given a boost both to automation and robotics in the enterprise. The global market for robots is expected to grow at a compound annual growth rate of around 26% to reach just under $210 billion by 2025, according to Statista. Deloitte anticipates that of the almost 1 million robots sold for business use in 2020, just over half were professional service robots, generating more than US$16 billion in revenue — 30% more than in 2019.

With the release of DeepRacer, developers can change the behavior of their cars, enabling the prototyping of new robotics apps. As Amazon notes, DeepRacer is essentially an Ubuntu-based computer powered by Robot Operating System, open source robotics middleware originated by Willow Garage and Stanford’s Artificial Intelligence Laboratory that provides low-level device control.

Sample apps

Six sample applications — Follow the Leader, Mapping, Off Road, RoboCat, DeepBlaster, and DeepDriver — are available to help with brainstorming, including one that creates visualizations of a home or office. As the DeepRacer community creates projects, Amazon says it’ll add them to the DeepRacer GitHub organization as well as feature them in future blogs.

“[W]e want to make it easier for developers of all skill levels to prototype new and interesting uses for their car. By making the AWS DeepRacer device software openly available, now anyone with the car and an idea has the ability to make it a reality,” Amazon wrote in a blog post. “Want to block other cars from overtaking it by deploying countermeasures? Want to deploy your own custom algorithm to make the car go faster from point A to B? You just need to dream it and code it.  We can’t wait to see the ideas you come up with, from new racing formats to new uses for the AWS DeepRacer.”

Amazon previously partnered with Udacity to offer machine learning courses and a scholarship based around DeepRacer. The ostensible goal was to educate students on the creation, training, and optimization of reinforcement learning models, or models that employ rewards to achieve goals. In a recent analysis, McKinsey pointed out that reinforcement learning can be applied to solve real-life problems beyond autonomous driving, including classification, continuous estimation, and clustering.