Posted On: Dec 1, 2019
Starting today, developers can take on their next machine learning challenge using AWS DeepRacer with the launch of multi-car racing and object avoidance capabilities in the AWS DeepRacer console. Customers can now build models for object avoidance and dual-car head-to-head races by experimenting with multiple sensor inputs and the latest reinforcement learning algorithms and neural network configurations. Developers can build reinforcement learning models ready to deploy to AWS DeepRacer Evo in the 2020 season of the AWS DeepRacer League.
Customers can start by adding stereo camera and LIDAR sensors to the virtual car in the new Garage section in the AWS DeepRacer console. The stereo camera adds depth perception to allow the car to detect objects in the road and be more responsive to its environment. The LIDAR sensor uses a continuously scanning laser to provide the reinforcement learning model with data about whether a car is fast approaching from behind.
With these new features the AWS DeepRacer League will expand and include new racing formats allowing developers to test these new found skills and race to win a trip to re:Invent 2020. In 2020 the League will add 8 more AWS Summit races in 5 new countries, and 18 virtual races which will be available in the AWS DeepRacer console. Customers will also have the ability to race in a league of their own by creating their own Community Race and compete with coworkers, friends or community members via the AWS DeepRacer console.