Which technologies did the organizers use in the self-driving cars in “Digital Race” 2018-2019? Let’s find out with FPT TechInsight.
Replacing the simply-assembled cars with basic appearance of last year’s “Digital Race”, this year the star of the show resembles a real racing car. From 1:10, the car size this year goes up to a 1:7 ratio in comparison with a standard self-driving car. With larger size and higher machine power, the car can reach a maximum speed of 100 kilometers per hour. The high speed appears to be quite a challenge for the competing teams, however, seeing that they need to carefully adjust the speed for a stable ride.
The highlight of this year’s car is that all tasks are processed through a motherboard. The higher investment means that the “brain” of this third season of “Digital Race” has the ability to solve various artificial intelligence (AI) algorithms such as Deep Learning and Machine learning, while cars in previous seasons struggled to do so.
For the car to be able to self-drive, competing teams had to utilize the image-processing technology on footages recorded by the front camera. On-car cameras can record in multiple degrees for a better function of this technology. This year, the competition had raised their bars, making the requirements more challenging but more practical, such as identifying passable and impassable obstacles, adding plants’ shades…
Integrated in the car is the standard braking and bumper system for model cars. The competing teams also needed to solve more problems compared to the previous year, regarding speed-control, reducing shock for cameras, all of which made the competition came closer to the real self-driving car.
At the semi-final, the teams had to overcome various challenges such as: route identification, sign processing, shadow processing (trees’ shade, human’s shadows…), laneless roads. The snow challenge, in particular, proves to be quite difficult to overcome, as image processor misread the snow’s whiteness as road markings, causing disruptions to the car operations. The results are calculated in the shortage finish time in a complete round (which must include the starting location, on-route destinations set by the organizers, and coming back to the starting point).
According to Mr. Le Hong Viet, the CTO of FPT, head of the competition’s organizer, having met all the requirements of the competition means that the participants had acquired all the basic knowledge needed to create a standard self-driving car.
During the competition, the cars self-operated without any directions and control, so the competitors are no different to viewers, with the exception of hardware malfunctions.
After 45 minutes of thrills and taunts, the semi-final results for the Southern regions were finally announced. The top 4 of the competition are SQ26 (Telecommunications University), CDSNTU2 (Nha Trang University), Dateh IT (Ho Chi Minh City University of Science), and LHU The Walkers (Lac Hong University), now officially the finalists of Digital Race 2018-2019.
The North’s semi-final will be hosted on 19:00, April 11th at Cau Giay Gymnasium, 35 Tran Quy Kien Street, Dich Vong, Cau Giay District, Hanoi. Readers who are interested in watching this round can register HERE.
Translator: Anh Thu