RMIT students rank 21st worldwide in AWS DeepRacer

RMIT students rank 21st worldwide in AWS DeepRacer

Team RMIT-badger of three RMIT School of Science and Technology students ranked 21st worldwide at the AWS DeepRacer hosted by the leading global cloud services provider - Amazon Web Services.

news-1-rmit-students-rank-21st-worldwide-in-aws-deepracer Team RMIT-badger of three RMIT School of Science and Technology students La Tran Hai Dang (pictured left), Hua Van Anh Khoa, and Le Dinh Danh Nhan (pictured right) ranked 21st worldwide at the AWS DeepRacer hosted by the leading global cloud services provider - Amazon Web Services.

Recruited by RMIT Centre of Digital Excellence (CODE), Bachelor of Information Technology students La Tran Hai Dang and Hua Van Anh Khoa, and Bachelor of Engineering (Robotics and Mechatronics) (Honours) student Le Dinh Danh Nhan, teamed up to compete in the AWS DeepRacer.

AWS DeepRacer provides a 3D racing simulator with a virtual race circuit where software developers of all skill levels [from beginner to expert] can participate in by programming a virtual race car to race autonomously on the circuit.

After only two weeks of competition, the team has been impressively ranked 21st worldwide, beating more than 2,000 other teams around the world, and is now qualified to compete at the AWS DeepRacer Pro Division. 

news-2-rmit-students-rank-21st-worldwide-in-aws-deepracer After only two weeks of competition, the team has been impressively ranked 21st worldwide, beating more than 2,000 other teams around the world, and is now qualified to compete at the AWS DeepRacer Pro Division.

The team was thrilled with the enormous opportunity given by such achievement.

“The ranking has opened doors for us to move on to the next challenging rounds,” La Tran Hai Dang said.

“While preparing for the competition, we were able to access much more new and useful knowledge about machine learning and deep learning which will be very helpful for us in the future.”

Le Dinh Danh Nhan emphasised that acquiring a lot of new knowledge, such as Deep Reinforcement Learning, PPO, sparse rewards, DRS, is a big support for his future.

“Controlling and creating my own fantasy world of robotics and virtual reality might no longer just be a dream for me,” Nhan said.

Opportunity comes with challenges. There were many obstacles along the way which the team had to overcome with.

“This is the world's first machine learning competition, so there is very little reference and materials that we can search in order to prepare for the competition,” Hua Van Anh Khoa shared.

“Machine learning and autonomous vehicles are also new fields in computer science where there is very little documentation or video tutorials available, and its practical applications are limited.

“We had to join a community to network with other competitors to exchange and share knowledge.”

Team RMIT-badger’s virtual race car raced autonomously on a virtual race circuit in a 3D racing simulator.

Khoa said that thanks to their strong background in programming as science and technology students, the team can easily learn new knowledge in the field but “it takes a lot of time to read and understand”.

Khoa believed it helped him gain an overview and experience in machine leaning - a potential and core field in the future of 4.0 revolution.

Now the team is racing against the clock to prepare for the coming rounds where they have to compete with professionals in the field of machine learning and autonomous vehicles.

“The competition does not require players to find a new way or spend time adjusting parameters to make the car run better. Instead, the players are asked to apply machine learning knowledge into practice,” Khoa shared about his realisation of the importance of machine learning algorithm in the competition.

“The team with the better algorithm will be advanced to the Pro Division.”

“Everything from knowledge to real examples and sample algorithms is ready for us to explore. It is just about how long it will take us to acquire knowledge and find a way to apply it to virtual cars in the next AWS DeepRacer.”

Story: Ha Hoang

  • CODE
  • Engineering

Related news