On September 9, the 2020 Yuelu summit was held in Changsha. At the special session of intelligent Internet connection, Su Kuifeng, general manager of Tencent’s automatic driving, delivered a speech entitled “digital twin promotes the evolution of automatic driving and intelligent transportation”. “Supported by real data, digital twin technology realizes seamless interworking between virtual world and physical world, provides a free test site for autonomous driving and intelligent transportation solutions, and enables people to foresee the infinite possibilities of future transportation,” he said
Since the founding of the people’s Republic of China, China’s transportation industry has developed rapidly and established a transportation network extending in all directions. At the same time, with the rapid development of economy, China has experienced the largest and fastest population urbanization in the world, as well as the explosive growth of car ownership, which has caused great pressure on the transportation system.
In the case of the traditional infrastructure dividend peaking and the scale of the transportation system is large enough, in order to meet people’s growing travel demand and bring better travel experience, it is necessary to use intelligent and digital technology to drive the landing of cutting-edge technologies such as automatic driving and vehicle road coordination, accelerate the intelligent upgrading of the transportation system and improve the transportation operation efficiency. This is not only an important part of China’s new infrastructure construction, but also an important part of achieving the strategic goal of strengthening the country through transportation proposed by the Party Central Committee and the State Council.
Since its establishment in 2016, Tencent autopilot has always taken software and services as its core business direction, integrated and utilized Tencent’s technical advantages in big data, cloud computing, artificial intelligence, etc., to help the development of automobile and transportation industry. “We have been insisting on consolidating the foundation of Industrial Development — cloud, simulation and high-precision map, which not only supports the development and operation of the automatic driving system, but also helps to build intelligent transportation through platform management such as data management and vehicle management.” Su Kuifeng said.
Tencent autopilot business architecture
Integrate multiple technologies to open up the virtual world and the real world
With regard to intelligent transportation, the future direction is relatively clear. In terms of individual travel, many domestic automobile enterprises and technology enterprises have invested heavily in the fields of automatic driving, Internet of vehicles and vehicle road coordination. In the field of public travel, the major Internet giants actively cooperate with relevant government departments and transportation service enterprises to explore various solutions from the perspective of the whole city operation, so as to realize the reasonable allocation of the overall road traffic resources and improve the travel efficiency.
However, whether it is personal travel, public travel, or the overall operation of urban transportation, China’s exploration of intelligent transportation is still in development. Most of the cutting-edge technology applications and solutions need countless tests and adjustments to truly achieve large-scale commercial use. This process requires huge time and money costs, and many tests are in reality It is difficult to carry out in China. All these are obstacles for China to move towards intelligent transportation and build a transportation power.
Tencent autopilot integrates the core capabilities of its cloud computing, high-precision map and virtual simulation platforms. With the help of game engine, three-dimensional reconstruction and other technologies, based on real traffic data, the digital twin world constructed by Tencent autopilot is a complete mapping of the real environment, and can freely switch time and space and change scenes according to the physical rules and operation logic of reality, which is a low cost Cost and high efficiency are the best place for future transportation.
Digital twin helps the development of automatic driving and intelligent transportation
In the field of autopilot, based on the real scene data, the high fidelity traffic scene can be recreate in the three-dimensional virtual environment, and the digital twin world can become the cloud test field of the autopilot vehicle. Traffic flow AI trained with a large number of real road data can generate highly realistic traffic flow models with reference to the logic of human behavior in the real world. Coupled with the underlying ability support of high-precision maps and the rendering of different lighting and weather environments by game engine, digital twin world can generously generate numerous test scenarios on the basis of real data for self-learning Full support for dynamic driving test.
In addition, the digital twin world, with the support of cloud capabilities, can run many test scenarios at the same time, and can randomly combine many test scenarios into city level traffic scenarios, in which thousands of automatic driving vehicles are deployed for 7 × 24-hour uninterrupted test, quickly locate the scenarios that the automatic driving algorithm can not handle well, and carry out targeted optimization.
The digital twin built by Tencent can carry out city level simulation and high concurrent automatic driving test. This ability can also be used to deduce intelligent transportation solutions for the future.
With the help of automatic driving target detection technology, pedestrian and various traffic elements are extracted from the real traffic image, and then integrated with the three-dimensional environment, the real road traffic situation can be seen in the virtual environment. Through the long-term collection of a large number of real traffic data, parameterization and modeling, the digital twin world can present the actual traffic situation, produce high authenticity traffic data, and provide decision-making basis for traffic scheduling.
In a city level traffic simulation scene, engineers can adjust the parameters, freely test the strategy simulation of bus first, traffic flow control and guidance, and intuitively evaluate its effect. In addition, it can demonstrate the application of v2x, vehicle road collaboration and other technologies, and it can also make traffic lights information-based and intelligent, so that they can switch freely according to traffic flow, or even switch different weather conditions at any time, evaluate the impact and formulate corresponding plans.
Data driven opens the fast lane to the future
In the digital era, everything can be described in the form of data, and parameterized and modeled according to the inherent laws. Tencent autopilot builds a digital twin world based on rich and diversified historical and real-time data. Through the simulation model, it operates the digital world with strong cloud computing ability and referring to the laws of reality, so as to open the “fast lane” leading to the future smart transportation and smart city life.
At present, more and more test fields and car companies have stepped onto this “fast lane”. At the end of last year, Tencent autopilot cooperated with the national intelligent networked vehicle (Changsha) test area to accurately reproduce the geographical panorama of Changsha test area in three-dimensional space by using virtual simulation technology, which is used to support safe and efficient intelligent networked vehicle simulation test and help the development of intelligent transportation in Hunan Province.
Digital twin technology is accelerating the implementation of intelligent transportation applications and promoting the strategic process of China’s transportation power. In the future, Tencent autopilot will always adhere to its own positioning, assist industrial development with technology and services, and jointly build “people-centered” intelligent transportation with partners.