Registration for the 2022 China Intelligent Connected Car Algorithm Challenge (CIAC) officially starts

/ Introduction /

##In response to the national science and technology leadership , innovation-driven development strategy, promote the development of intelligent connected automobile technology and industry, accelerate the training of talents in this field, enhance the independent controllability of China's industrial software, while focusing on the key technologies and major challenges faced by the development of the intelligent connected automobile industry, 中国Artificial Intelligence Society, China Automotive Engineering Society, and China Productivity Promotion Center Association jointly sponsors the 2022 China Intelligent Connected Car Algorithm Challenge (CIAC 2022) in conjunction with key national universities, scientific research institutions, and technology companies. The competition is organized by The Intelligent Driving Professional Committee of the Chinese Society for Artificial Intelligence, the National Intelligent and Connected Vehicle Innovation Center, the Intelligent Transportation Branch of the China Society of Automotive Engineers, the Automotive Working Committee of the China Association of Productivity Promotion Centers, and the Automotive Technology Education Branch of the China Society of Automotive Engineers are jointly organized by professional organizations.

China Intelligent Connected Vehicle Algorithm Challenge will attract a wide range of students from well-known universities at home and abroad, technicians from autonomous vehicle companies, and researchers from scientific research institutes. Participate. As one of the most authoritative and influential competitions in the field of autonomous driving in China, the challenge is held every year. People from all walks of life at home and abroad are welcome to actively pay attention and participate.

/ Competing time /

Registration time: 2022.9.30 – 2022.10.24

Qualifier: 2022.10.25 – 2022.11.4

Final: 2022.11.5 - 2022.11.15

##Awards Ceremony: 2022.11.22 - 2022.11.24

##/ Competition items/​

Perception Competition Question 1: Visual Detection of Vehicles on Highways and Urban Roads

This The competition question is based on the forward-looking image data collected by natural driving. In highway and urban road scenes, contestants need to detect motorized and non-motorized vehicles in the scene. The occlusion or truncation ratio is less than 60%, and the short side pixels are greater than 12 All vehicles need to be detected and the two-dimensional bounding box (BoundingBox) of the vehicle needs to be output. When the vehicle is occluded, the bounding box needs to include the occluded portion of the vehicle. When the vehicle is cut off (part of the vehicle is outside the lens), the bounding box only needs to include the portion of the vehicle that is visible within the image.

Perception competition question 2: Visual detection of corner points and edge points of parking space lines in parking lot scenes

This competition question is based on the bird's-eye view collected by the high-definition fisheye surround view camera and processed by AVM to form a bird's-eye view. The corner point analysis of parking spaces with two or more corner points in the parking lot parking scene is performed. Detect edge points that are blocked or cut off by objects. Parking space types include vertical parking spaces, horizontal parking spaces and inclined parking spaces, and output the pixel coordinates of corner points and edge points of the parking spaces.

Perception Competition Question 3: Point Cloud Traffic Participant Detection on Highways and Urban Roads

This competition question is based on the natural driving data collected by 128-line mechanical lidar. It evaluates the detection performance of point cloud traffic participants in highway and urban road scenes. It requires that the number of point cloud points is greater than 5. The traffic participants at the selected point are detected, the detection range is 100 meters around the vehicle, and the type and posture of the traffic participants are output.

Simulation Competition Question 1: Formula Student Scenario

This competition scenario refers to the "High-speed Tracking Test" event in the "China Student Driverless Formula Competition", which mainly tests the stability and response speed of the path planning and path tracking algorithms. The high-speed tracking test track is a closed-loop track that includes straight lines, fixed-radius turns, and hairpin turns. It combines the characteristics of various test performances such as acceleration, braking, and steering

Simulation Competition Question 2: Highway Scene

The purpose of this competition is to test the automatic driving system algorithm in Performance under complex high-speed working conditions. The site is a multi-functional ring test site composed of straight roads, large curvature curves, and entry/exit ramps. During the operation of the main vehicle, the surrounding traffic vehicles will actively interact with it in an adversarial manner, including the following typical working conditions: rapid acceleration, rapid deceleration, close cut, multi-vehicle confrontation, and ramp merge/exit rush, etc. Vehicles need to enter from the ramp entrance, then go around the ring test site for one week, and finally exit from the ramp exit.

Simulation Competition Question 3: Urban Intersection Scene

The purpose of this competition is to test the performance of the autonomous driving system algorithm in complex scenarios at urban intersections. The competition questions include the following 4 sub-questions: go straight at the intersection, turn left at the intersection, make a U-turn at the intersection, turn right at the intersection and meet pedestrians. During the competition, cars will be randomly sent in all directions at the intersection to simulate complex interactive behaviors in the real world. The starting point of the race is in front of the intersection, and vehicles need to pass the intersection as required to complete the race.

Simulation competition question 4: Parking lot parking scene

The purpose of this competition is to test the performance of the automatic parking algorithm under standard working conditions and non-standard parking conditions with Chinese characteristics. On the basis of the national standard GB/T41630-2022, the competition content combines China's regional characteristics and the industry's cutting-edge technology needs, adding non-standard parking spaces such as small parking spaces and small controllable area distances, increasing the complexity of the competition scene.

/ prize settings/

The challenge has first, second and third prizes

The winners will receive a certificate of honor issued by the competition organizing committee. You will also have the opportunity to receive bonuses, letters of recommendation, extra points for the Chinese Student Formula Driverless Competition (FSAC), etc. Details of the awards will be announced later, so stay tuned!

##/ Registration Channel/

The QR code for the challenge registration channel is as follows

You can also log in to the CICV automotive developer community (developer.china-icv.cn) to register

Communicate the QR code of this competition group to get the latest information

##Perception Communication Group     Simulation Communication Group​

#///

(1) Participants: The competition is recruiting teams from all over the world, regardless of age or nationality. Employees from universities, research institutes, companies, etc. can log in to the CICV automotive developer community to register for the competition;

(2) Developers register as a team, with 1 to 5 members. When registering, all members must provide basic personal information and register through the registration platform. For certification, the developer should ensure the authenticity of the information provided, and the organizer will keep the content involving personal privacy confidential;

(3) The development team must Have a suitable team name.

##/ Work requirements/

Perception Class

(1) Provide Docker image and Dockerfile file: Docker image must contain the required environment and algorithm Source code, etc., and name it with "task number_team name.tar", such as task1_zhangsan.tar;

(2) Provide detailed documentation: include The overall description of the task algorithm, solution ideas, architectural design, operation instruction instructions, etc., the file format is pdf;

(3) Other related supporting materials, such as: auxiliary Display design, plan materials, demonstration video, etc. (not mandatory);

(4) All the above materials are placed in the "Task Serial Number_Team Name_Version Number" " folder, and then compress it into "task serial number_team name_version number.tar" for submission, such as task1_zhangsan_v1.tar, the version number starts from v1;

(5) Original work: The work must ensure originality, do not violate any relevant laws and regulations of the People's Republic of China, and do not infringe any third party's intellectual property rights or other rights. Once discovered or submitted and verified by the rights holder, the organizer will cancel its participation Qualifications and results will be dealt with seriously;

(6) Reproduction and verification of works: Contestants need to cooperate with the organizer to verify the validity and authenticity of their works. Verify, and at the same time check the correctness of the submitted work by yourself, and submit it after confirming that it is correct. The organizer is not responsible for making changes and adjustments to the competition works.

(7) Each team can submit up to 3 times during the submission period, and the result of the last submission shall prevail.

Simulation class

(1) Submission based on Matlab /Algorithm code in Simulink, C, Python language;

(2) Please name the compressed package in the format of [Problem Number-Team Name], for example [Simulation Category Question 1 - Challenger Team];

# (3) Original work: The work must ensure originality and not violate any relevant laws and regulations of the People's Republic of China. No infringement of any third party's intellectual property rights or other rights. Once discovered or submitted and verified by the rights holder, the organizer will cancel his participation qualifications and results and deal with it seriously;

(4) Work reproduction and verification: Contestants need to cooperate with the organizer to verify the validity and authenticity of the work, and at the same time check the correctness of the submitted work by themselves, and confirm it before submitting , the organizer is not responsible for changes and adjustments to the competition works.

##/ Judging Rules/

(1) The organizer of perception competitions will provide relevant data sets; the organizer of simulation competitions will not provide batch data of the competition questions, only task sample data;

(2) All works submitted by qualified teams before the required date will be included in the review. The organizer is not responsible for any damage, loss, submission delay, etc. caused by computer, Internet, or mobile network failures;

(3) It is prohibited to plagiarize other people's works or exchange answers during the competition. Once found, the results of the competition will be canceled and serious punishment will be taken.

/ organization/

organizer

China Artificial Intelligence Society

China Automotive Engineering Society

China Productivity Promotion Center Association

# #organizer

Intelligent Driving Professional Committee of China Artificial Intelligence Society

National Intelligence Network United Automotive Innovation Center

Intelligent Transportation Branch of China Society of Automotive Engineers

China Productivity Promotion Automotive Working Committee of the Central Association

Automotive Technology Education Branch of China Society of Automotive Engineering

The above is the detailed content of Registration for the 2022 China Intelligent Connected Car Algorithm Challenge (CIAC) officially starts. For more information, please follow other related articles on the PHP Chinese website!