Hello, welcome toChina Autonomous Driving Testing & Validation Technology Innovation Conference 2021!

L3 Autonomous Driving Road and Virtual Test

Release Date:2019-07-20

Shanghai, China, June 27, 2019 - In “ The 2nd Annual China Automated Driving Test Validation Technology Innovation Conference “ , Professor Wang Jian from the School of Computer Science of Jilin University, participated this conference and delivered a keynote speech.



In the conference, Mr. Wang Jian first explained the autopilot scene from the definition, composition and key features of the scene. The scene research is the key technology of intelligent driving technology and product development. The scene library is an important part of the whole test. The input in the entire test session gives us the evaluation in the test and judges a position of the scene. The driving environment is infinitely rich, extremely complex, unpredictable and inexhaustible. The composition of the scene is to divide the complex driving environment into two major parts: static features and dynamic features. Static features include road sites, traffic and facilities, and dynamic features include Traffic, weather, etc. The static features are extracted, and the virtual algorithm is used to generate dynamic features. The combination of the two is the basic composition of the scene.


A scene is a finite mapping of an infinite world, whether it is a static feature or a finite mapping of dynamic features, and then a measure of coverage, coverage of dangerous scenes, and accuracy of testing. These two are the two results that are reasoned out by the automatic generation. The generation of the scene library is to cover the infinitely rich and extremely complex driving environment through limited mapping, and finally generate the scene library. The specific steps are as follows: firstly, a record of the driving environment is mapped to the network, and by learning the real scene features, a new scene is derived, and the scene library indicators that need to be stored are extracted, and finally abstracted into a driving situation. And driving occasions.






The three key elements of the scene's impact on autonomous driving are driving situations, environmental impacts, and driving scenarios. Driving conditions such as highways, country roads, urban roads, etc., these changes are not very strong, generally do not change after the selection; environmental impacts such as roads, traffic, pedestrians, weather, etc., these are the most complicated situations. The key to environmental impact is environmental sensing perception, lidar and millimeter wave radar, cameras, positioning systems, V2X communication equipment, we need to understand which factors affect the sensor and store these conditions in the scene library. Driving situation: The driving situation is an important external factor of the scene. The driving situation is divided into the following three types: 1. The driving tasks of the vehicle are: lane change, overtaking, U-turn, turning, etc.; 2. The driving speed of the vehicle is as follows: acceleration, deceleration, etc. 3, the driving mode of the vehicle is conservative, radical or normal.



Afterwards, Mr. Wang Jian introduced the key technologies of digital virtual scene construction through data source and simulation scene construction: data source: many sources of scene data, including natural driving development road scene, closed demonstration area scene, accident reconstruction, simulation scene Different forms or platforms to build data sources. Scene data from different forms or platforms have different characteristics and applicability, and good complementarity. These four scenarios are complementary. Construction of simulation scenes: The construction of simulation scenes includes laboratory simulation scenes, driving simulator acquisition scenes and digital virtual scenes. They have good randomness and coverage, and are especially suitable for constructing a huge number of unknown, dangerous and extreme driving scenarios. Annotation can be generated automatically to support deep learning.




Finally, Mr. Wang Jian showed the guests how to combine the laboratory simulation scenario construction and the scene library with the test tools. Through the practice test cases and test data, the guests showed the L3's autopilot test in detail, which gave great inspiration to the car manufacturers and laid the foundation for the early realization of autonomous driving.


Personal profile of the guest speaker: Wang Jian, professor of computer science at Jilin University, doctoral tutor, special senior expert of China Automotive Engineering Research Institute, and senior expert of Qiming Information Technology Co., Ltd. He has worked as a doctoral student, postdoctoral fellow, and visiting scholar at the University of British Columbia in Canada, the University of Innsbruck in Austria, the National Institute of Information and Automation in France, and Hanyang University in Korea. The main research areas are communication protocols, MEC applications, simulation and testing of intelligent networked vehicles. He is currently the Secretary-General of the International Parallel Driving Alliance, the Standing Committee Member of the Parallel Intelligent Professional Committee of the China Society of Automation, the Communication Committee of the China Intelligent Transportation Industry Alliance, the Member of the China Auto-E Alliance, and the Youth Committee of the Knowledge Engineering and Distribution Intelligence Committee of the China Artificial Intelligence Society. Member of the editorial board of Vehicular Telematics and Infotainment Systems, former co-chair of the 29th IEEE Intelligent Vehicles Symposium Publishing Committee, IEEE VTC, CV2N, VTHWN, SAE 2017 ICVS, IWCMC 2017 and other international conference technical committee members. In recent years, as the project leader, he has undertaken the National Natural Science Foundation of China, youth projects, international cooperation and exchange, special fund for China Postdoctoral Fund, Doctoral Fund of the Ministry of Education, Jilin Province Development Plan Key Project, Jilin Province International Cooperation Project, Jilin More than 10 vertical projects such as the Provincial Youth Fund, 1 second prize for national education achievement in higher education, 1 first prize for Jilin Science and Technology Progress Award, 1 first prize for teaching achievements in Jilin Province, China National Commercial Federation The second prize of the Business Technology Progress Award was 2 times. The first responsible author published 42 papers in the International SCI Search Journals such as IEEE Transaction on ITS, IEEE Transaction on IV, Scientific Reports, Computer Networks, Computer Communications, and authorized 7 invention patents. There are 4 software copyrights, and one monograph in Chinese and English.


Organizer: Shanghai Locka Automotive Technology Co., Ltd. is a technical consulting company engaged in the development of technology in the field of automotive technology, technology transfer, technical consulting exhibition services, conference services and other business development. The company provides industry information, business innovation development solutions, market research, business cooperation and network development platforms, personal career development, investment and financing consulting services to senior decision-makers at leading domestic and foreign companies (mainly Fortune 500 companies).


"China Automated Driving Test and Verification Technology Innovation Forum 2019 (CADT2019)"

  • 电话咨询
  • +862122306692
  • 15021948198