Covering Disruptive Technology Powering Business in The Digital Age

Writen by: Izzat Najmi, Journalist, AOPG.

Amazon introduced AWS DeepRacer back in 2018 as a light-hearted means of introducing programmers to machine-learning. Since then, the company has made adjustments, such as adding DeepRacer tournaments, and now they have revealed yet another improvement. The company is releasing the code it developed to control these tiny vehicles as open-source.

During the keynote of the AWS Public Sector Summit, Elsie Tan said “At AWS, we believe machine-learning will be the most transformative technology of our generation. By putting machine-learning technology in the hands of every developer, AWS is committed to enabling our customers to build new, innovative solutions that improve lives and protect our planet. AWS DeepRacer makes these complex concepts easy to understand, apply, and most importantly, FUN! This year, we have hosted AWS DeepRacer events with our customers like the Institute of Technical Education, Central Provident Fund Board, and InfocommMedia Development Authority, to help them demystify some of these technical concepts.

The DeepRacer automobile is fundamentally an Ubuntu Linux–powered minicomputer, with the open-source Robot Operating System (ROS) layered on top. The business hopes that allowing programmers wider access to the device’s code will lead to more innovative uses of the car by letting them alter the vehicle’s typical operation.

One of the AWS DeepRacer cars in action

Since the AWS DeepRacer device code is now publicly available, you may easily adjust the way your car reacts to the circuit. Do you wish to prevent other vehicles from overtaking it by erecting barriers? Do you wish to implement a special algorithm of your own design to hasten the journey from A to B? It’s as simple as “imagine it and code it,” Mike Miller, General Manager, AWS AI Devices at Amazon Web Services (AWS) said during the AWS Public Sector Summit.

Since launching the vehicles in 2018, DeepRacer has created both live leagues where drivers compete against one another and online competitions. In fact, just last month, the business restructured the leagues in an effort to get more people interested in the technology participating. Providing developers with the freedom to make this their own could spark even more interest, as it would allow for entirely new uses for the cars.

The AWS DeepRacer is a remote-controlled, 1/18th size race car used for testing Reinforcement Learning (RL) models, either virtually in the AWS DeepRacer gaming console or in real life at AWS and customer events. Even if you do not have any experience with machine-learning, you can still use AWS DeepRacer as the user-friendly application helps you go from zero to hero. Participating in the AWS DeepRacer League is a great way to gain practical experience with machine-learning while also having fun studying RL using AWS DeepRacer.

AWS also introduced the virtual time trial booth where the AI system will calculate and compute your style of driving to the existing teams (participating in the AWS DeepRacer League) and suggest it based on four techniques:

• Centre Line Following.
• Steering Aggression.
• Average Speed.
• Speed Consistency.

Najmi, AOPG’s Journalist, testing the Virtual Time Trial Booth during the AWS Public Sector Summit

Get started with six example projects and access the project documentation on GitHub and the AWS DeepRacer Open-Source page to modify your car’s software.

(0)(0)