On September 10th, Tencent 2020 Global Digital Ecology Conference smart transportation forum was held on the cloud. In the context of new infrastructure in full swing, transportation, as the "main force" of new infrastructure, has also become the focus of hot discussion among all parties in the industry. In this intelligent transportation forum, Tencent released a new upgraded We Transport scheme.
In terms of how to promote the evolution of autonomous driving and intelligent transportation through digital twin technology and help the construction of intelligent transportation technology platform, Su Kuifeng, general manager of Tencent Autonomous Driving, introduced: "We have invested heavily in the research and development of core technologies such as fusion perception, efficient map construction, multi-sensor fusion positioning, system decision making, path planning and system control, and combined virtual simulation, data cloud and high-precision map cloud platforms to build core capabilities in autonomous driving and smart transportation."
From virtual simulation to data center: Helping the construction of smart transportation digital twin platform
Tencent's self-driving team started early in virtual simulation. In 2018, Tencent released TAD Sim, an autonomous driving simulation system, and an upgraded version, TAD Sim2.0, in the first half of 2020. As a virtual simulation platform driven by game technology and real data in the industry, TAD Sim has been put into use in many intelligent network test bases, domestic leading automobile enterprises and autonomous driving research and development institutions in the fields of authenticity, comprehensiveness, visualization, lightweight and standardization. In September 2020, Tencent, together with the National Automobile Quality Supervision and Inspection Center (Xiangyang), started the construction of digital twin autonomous driving evaluation standards.
The superposition of new technologies has brought about the effect of "1+1>2". The combination of virtual simulation technology and cutting-edge technologies such as high-precision map, big data, AI and cloud computing not only promotes the popularization of the application of virtual simulation of autonomous driving, but also gradually extends to new application scenarios such as traffic flow simulation with virtual-real combination, urban traffic simulation and traffic simulation prediction in a larger scope. Build a digital twin platform.
Tencent's digital twin platform models virtual worlds as best it can, based on a lot of real data from the real world. A large amount of traffic data, such as the number of lanes, slope, traffic flow, and location of traffic lights, will be collected every day by sensors, roadside cameras, and mobile smart devices such as mobile phones. The collected data is transmitted to the cloud data management center through the communication network for data classification, management, cleaning, labeling and other work to form effective data resources; Then the intelligent engine will analyze, mining, retrieval, visualization and other processing of the data, and the data can be provided to the application layer in the form of flexible and convenient interfaces such as API/SDK.
Finally, the digital twin platform built with massive real data as "raw materials" will build a model of the physical world in the virtual space and realize the one-to-one mapping between the digital twin environment and the real environment. However, the difference lies in that the world of digital twin can break through the limitations of time and space, deduce the future urban traffic operation driven by real data, and explore the solutions of smart city planning and traffic operation management.
From technology platform to visual operation: Help users happy travel
The construction of digital twin technology platform is only the first step in the construction of smart transportation and smart city. Based on the digital twin technology, traffic flow AI trained on the basis of a large number of real traffic flow data generates the traffic flow combining virtual-real with high degree of reality, so as to forecast, deduce and verify the urban traffic system and comprehensively help the evolution of intelligent transportation.
"First, Tencent uses high-precision maps as basic data, combines all kinds of real environmental data, 3D reconstruction and game engines to achieve high-precision restoration of static scenes, and can render different lighting conditions and weather scenes; Then, through game engines, vehicle dynamics models, sensors and other dynamic models, the virtual scene has the same physical rules as the real scene. Finally, based on real-time traffic images of cities, Tencent uses object detection technology to extract road and vehicle targets, parameterize and model them, and integrate them into a three-dimensional environment." Su Kuifeng introduced.
It is understood that, based on Tencent cloud computing capability, high-precision map and positioning technology, three-dimensional reconstruction and other technical support, in the process of platform operation, traffic flow simulation can not only be realized in the micro scene, but also cover the middle simulation of key areas, and even large-scale lane-level simulation of city-level road network. At the same time, through historical data analysis and model inference, the digital twin platform can freely simulate various weather conditions, assess the impact of weather conditions on traffic operation, and rehearsing countermeasures. The traffic lights will be changed from fixed configuration to electronic intelligent police, which can conduct real-time intelligent scheduling according to road conditions and even make real-time prediction based on data, so as to make traffic more accurate and efficient. The roadside equipment will be intelligent, the vehicle and road cooperation based on 5G communication technology will be promoted, and the executable plan will be determined.
In terms of presentation mode, Tencent digital Twin platform makes use of superior game technology to visually present traffic flow data, form real-time traffic heat map, and visually present the overall traffic state. It supports real-time dynamic optimization of traffic intersection timing and simulation of bus advance strategy. Evaluation of roads and other infrastructure; Integrated traffic management and scheduling, congestion prediction and real-time navigation push and other services, help urban traffic governance to become digital, intelligent direction.
The smart transportation upgrade brought by the new infrastructure is centered on data elements, so that each link and each terminal in the traffic can realize the flow of data, so as to better play the value of production factors and drive the optimization of traffic construction, management, operation and service. Digital twin platform is exactly the infrastructure that enables data to play a greater value. Tencent hopes to unite all parties in the industrial chain to build an open and cooperative ecology, and finally realize more efficient and convenient urban transportation, bringing more pleasant travel experience to people.
