Chongqing Auto Enterprise Wows Audiences in Strongest Car Brain Challenge

After a night of rest for participants, the 2021 i-VISTA Grand Challenge resumed the competitions. On August 20, nine teams participated in the Strongest Car Brain Challenge and five in the online Virtual Simulation Challenge. In the end, the first and second teams from Chongqing Jinkang SERES New Energy Vehicle Design Institute Co., Ltd. won the special prize and first prize in the enterprise group of the Strongest Car Brain Challenge with wonderful performances; the team dispatched by the Beijing Institute of Technology (BIT) won the Virtual Simulation Challenge.

The first and second teams from Chongqing Jinkang SERES New Energy Vehicle Design Institute Co., Ltd. won the special prize and first prize in the enterprise group of the Strongest Car Brain Challenge with wonderful performances. (Photo provided to iChongqing)

The Strongest Car Brain Challenge mainly tests the perceptual and processing abilities of L4-L5 autonomous vehicles in comprehensive urban road conditions. The competition was held at a closed section of the track from Pufu Avenue to Xietong Avenue in Chongqing Liangjiang Collaborative Innovation Zone, with a length of about 3.2 kilometers. A total of 12 scenarios were set, including T-shaped intersections, pedestrians crossing the road, vehicle interference, lane merging, traffic lights, etc. There were quite a few extreme and dangerous scenarios such as "pedestrians suddenly crossing the road", which greatly enhanced the difficulty of the competition.

The competition officially began at 9 a.m. as the referee team, scenario team, conference team, and other staff took their positions. The Jinglong-Tianyan team from Beijing Union University departed first. The vehicle, equipped with sensing equipment on the roof, looks futuristic. After leaving the run-up area, the vehicle ran automatically at a limited speed of 60 km/h. According to experts from the event's organizing committee, in addition to a referee, there was a safety officer in each vehicle; once the vehicle was out of control and might endanger the driving safety, the officer would intervene to control the vehicle but the manual operation to turn the steering wheel or step on the brake would cause deduction of points.

At the scene, the reporter saw that when a vehicle was passing through a crosswalk, a dummy suddenly rushed to cross the road from the right side. It simulated exactly the scene of pedestrians crossing the road illegally. Fortunately, the vehicle was "smart" enough to perceive the dummy and respond quickly, stopping in front of the crosswalk. After this vehicle continued running for some distance, some pre-arranged vehicles came out for interference. It was also one of the scenarios that the referee team had predetermined. In the process of obstacle avoidance, one side of the racing vehicle overrode a solid line on the road. The referee immediately judged that points should be deducted for the behavior and the longer the behavior lasts, the more points should be deducted according to the rule.

Cars in the competition. (Photo provided to iChongqing)

"Intelligent vehicles run on smart roads, and vehicles, roads and humans form a smart traffic network. It aims to test whether autonomous vehicles can calmly respond to various complex scenarios with the help of network communication and collaborative perception." In this competition, the daily use scenarios of vehicles were accurately restored. In order to increase the difficulty, even a number of random scenarios were set, including random changes of traffic lights in signalized intersection conditions and random appearance of a pedestrian in the scenario that a vehicle should give way to the pedestrian before turning right and avoiding the pedestrian crossing a road. These variables make the race more challenging and test the vehicle's capability of immediate perception, decision-making, and control, according to a relevant expert of CAERI Intelligent Connected Technology Co., Ltd., the organizer of the event.

In the hot sun, all the teams did their utmost. With the last team off the track, the Strongest Car Brain Challenge came to a successful end. The two teams from Jinkang SERES representing Chongqing's auto companies won the first two places in the enterprise group. The Jinglong-Tianyan team and the Jinglong-Tianying team from Beijing Union University secured the special prize and the first prize in the university group. The first Bi Sheng team from Chongqing University of Posts and Telecommunications clinched the award of excellence.

While nine teams battled for the honor of the Strongest Car Brain, five teams competed online in the Virtual Simulation Challenge. The simulation scenarios of this competition include four typical road structures, namely, intersection, T-entrance, on-ramp and off-ramp, and river area, and test the ability of the autonomous driving system to make decisions and control algorithms in the preset scenarios of the virtual traffic environment.

This year's Virtual Simulation Challenge introduced cloud platform for the first time. (Photo provided to iChongqing)

It is worth mentioning that this year's Virtual Simulation Challenge introduced a cloud platform for the first time. PanoSim, the simulation engine designated by the organizer, carried out in-depth technical optimization and cooperation with Tencent Cloud for Autonomous Driving. On Tencent Cloud for Autonomous Driving, PanoSim simulation engine can realize the efficient operation of scenarios and establish a perfect algorithm evaluation system in many dimensions, such as laws and regulations, safety, comfort, and passing efficiency. After competitors upload the algorithm to the cloud, online judging can be automatically realized and results can be quickly generated, which ensures openness and transparency throughout the whole process.

The virtual fight proved the real competencies of the teams. In the end, the special prize, the first prize and the award of excellence were won by the team from Beijing Institute of Technology (BIT), the team from the University of Electronic Science and Technology, and the J1501 team from Chang'an University respectively.