JBRRussia1, the team of young developers from St. Petersburg, took first place in one of the disciplines of the competition on the use of artificial intelligence for self-driving cars. It took place in Vancouver (Canada) at the 33rd conference of Neural Information Processing Systems (NeurIPS), the largest and most prestigious event in the field of artificial intelligence and machine learning.
Competitions used the Duckietown platform, a training ground for testing algorithms of autonomous vehicles, which is a reduced model of the urban transport environment, including road markings, vehicles, traffic lights, and buildings.
Competitors from Russia, Hungary, Canada, Switzerland, the USA, Germany, India, Egypt, and Colombia presented the algorithms they developed for lane following and movement around the city - they tested each solution in a simulator and on a real bot. The grades included such criteria as the robot operation time, the distance traveled, and the degree of deviation from the center of the lane.
In the Lane Following Challenge, the Russian team presented a solution that showed the best result both in a simulator and on a real bot. According to the results of the competition, the Russian young researchers took first place, beating their main competitors - the phmarm team from Canada.
The Russian team included Konstantin Chaika, postgraduate student of the Department of Software Engineering and Computer Applications of ETU “LETI,” Nikita Sazanovich, student of the Higher School of Economics, and Maxim Kuzmin, a graduate of the St. Petersburg Computer Science Center. The team leaders are Kirill Krinkin, Head of the Department of Software Engineering and Computer Applications and Alexey Shpilman, Senior Lecturer at the Informatics Department of the Higher School of Economics Campus in St. Petersburg. Association of scientific groups JetBrains Research supported the preparation for the competition.