Launched a year ago, the AWS DeepRacer League is pitched as a global autonomous racing league for developers. The goal of the league is to build and race fully autonomous 1/18th scale race cars using code and simulation.
The original DeepRacer fully autonomous race car featured an Intel Atom processor, a 4 megapixel camera with 1080p resolution, fast WiFi, multiple USB ports, and about two hours of battery life.
The new DeepRacer Evo car adds more ways to learn about reinforcement learning with a new stereo camera and a Light Detection and Ranging (LIDAR) sensor, which helps the car to detect and respond to obstacles.
AWS is also announcing two new race types focused on object detection and avoidance. The goal is to utilize sensors to detect and avoid obstacles as the car goes head-to-head against another DeepRacer on the same track.
Meanwhile, AWS is also expanding the DeepRacer League with 8 additional races in 5 countries, 18 additional virtual races and a race at the upcoming re:MARS conference in 2020.
The DeepRacer updates are coming out in advance of AWS re:Invent, which will feature the final qualifying race and championship cup race for the 2019 DeepRacer League.