According to Mike Miller, General Manager of AWS Artificial Intelligence Devices, the previously proposed DeepRacer competition uses a 1/18 scale self-driving car, combined with AWS's machine learning and other technical resources, to allow users to create plans that can accelerate around the track at even faster speeds. The hope behind this is that more people will become familiar with artificial intelligence technologies like deep learning.
Mike Miller said that the DeepRacer competition program has attracted many students, businesses and other groups to participate, and has used AWS's machine learning resources to build various artificial intelligence technology applications, allowing mini racing cars to shorten the time required to complete a single lap on the track through computer vision and deep learning.
Just like a real self-driving car, AWS's 1/18th scale autonomous racing car uses LiDAR components, video cameras corresponding to computer vision, as well as sensors such as accelerometers and gyroscopes to determine current driving conditions and complete forward image recognition, allowing the vehicle to automatically determine how to complete a single lap using the optimal route while achieving the shortest completion time.
Through the implementation of the DeepRacer project, participants can more easily understand the actual operating principles of artificial intelligence technologies such as machine learning, and thus apply them to more application areas, such as automated engineering and computer vision analysis. At the same time, it can also expand its application to application scenarios such as smart manufacturing, smart healthcare, Internet of Vehicles, and smart cities, and enable the popularization and application of artificial intelligence technologies at a faster pace.
Behind the DeepRacer project, AWS also provides application resources such as DeepLens and DeepComposer. Users can even create application models through tools such as AWS SageMaker. During this year's Re:Invent 2021, AWS also proposed a coding-free design method, allowing users to only conceive the actual application model, and then through intuitive image-guided operations, they can quickly complete the artificial intelligence technology application settings through simple clicks, dragging, etc.
As for whether the technology behind the DeepRacer competition program has its limitations, Mike Miller believes that it mainly depends on how users can use their creativity to create actual content. He also emphasized that the more important thing about this program is to allow more people to access and learn the principles of artificial intelligence technology, while quickly connecting to a wider range of artificial intelligence application technologies.
