With the help of intelligent construction equipment, transportation engineering construction operations are more precise and reliable in quality; Equipped with intelligent transportation system, accidents are reduced and traffic is smoother; Relying on intelligent design algorithms, routes are more reasonable and travel is greener... "Artificial intelligence+transportation" is changing the way of production and life. Recently, the Ministry of Transport and seven other departments jointly issued the "Implementation Opinions on 'Artificial Intelligence+Transportation'", deploying 16 specific tasks around technology supply, scenario empowerment, and other four aspects. It is proposed that by 2030, the intelligent comprehensive three-dimensional transportation network will be fully promoted, with key core technologies being independently controllable, and the overall level ranking among the top in the world. The deep integration of artificial intelligence and transportation is based on solid technological support, which can help improve the efficiency of demonstration channels by about 20%. The opinions focus on three major directions: application technology research and development, intelligent product innovation, and comprehensive transportation model construction, and increase the supply of key technologies. "The emphasis on increasing the supply of key technologies is aimed at opening up the innovation chain from technology research and development to product application." Wang Yunpeng, an academician of the CAE Member and president of Beijing University of Aeronautics and Astronautics, said that in particular, he clearly proposed to build a large model of comprehensive transportation, and to provide a unified model capability support for the intelligent transformation of the industry by building a "transportation brain" covering various high-quality data sets, algorithm libraries and tool chains, so as to promote technology sharing and collaborative innovation. In recent years, the construction of large-scale transportation models has accelerated. At the national level, in August of this year, the Transportation Big Model Innovation and Industry Alliance was established under the guidance of the Ministry of Transport. Currently, it has gathered more than 50 industry leading enterprises, artificial intelligence companies, and related universities and institutions, and sorted out 860 typical application scenarios of artificial intelligence. At the enterprise level, China Logistics Group has released the "Flow Cloud" large model, which can be applied to more than 40 segmented scenarios such as multimodal transport and warehouse scheduling; The "smart hub" built based on Baidu's transportation model has been implemented in more than 10 cities, connecting over a million vehicles equipped with L2 level assisted driving functions. Various demonstration regions across the country are carrying out applications such as road network operation monitoring and early warning based on large models, active control of high traffic sections and trunk lines, and one-stop travel services. ”The relevant person in charge of the Ministry of Transport introduced that through the intelligent expansion of trunk channels and the construction of multi-level refined traffic control models, the efficiency of demonstration channels can be improved by about 20%, and the efficiency of emergency response to emergencies can be improved by about 30%. Artificial intelligence creates a solid digital foundation for transportation infrastructure through big data analysis, high-precision modeling, and other methods. Data shows that in 20 demonstration areas for digital transformation and upgrading of highway and waterway transportation infrastructure, the total length of demonstration channels exceeds 60000 kilometers, including about 54000 kilometers of highways and 7500 kilometers of waterways, basically covering the main framework of the national comprehensive three-dimensional transportation network. In the opinion of Li Hui, an academician of the CAS Member and a professor of Harbin Institute of Technology, the construction of the "artificial intelligence+transportation" digital base will effectively support the digital level of transportation infrastructure to achieve energy level leap, and provide a scientific basis for transportation planning, construction and management. The key to the deep integration of artificial intelligence and transportation in the deployment of combination assisted driving and other fields lies in the rich application scenarios. The opinion system has deployed intelligent application scenarios in seven key areas, including combination assisted driving, intelligent railways, and smart shipping, to accelerate innovation scenario empowerment. These scenarios almost cover all sub sectors of transportation, providing a richer testing ground and application space for new technologies and products, "said Wang Yunpeng. Let's first look at water transportation. 52 automated terminals have been built, and the independently developed fully automated container terminal intelligent operating system has been applied in more than 10 terminals at home and abroad. The electronic map of the waterway has been successfully applied to the main and branch waterways of the Yangtze River... The intelligent transformation of ports, waterways, and ships has accelerated, effectively releasing water transport efficiency. China's inland waterway transportation is facing problems such as overall low waterway grades and a small, scattered, and mixed shipping market. "To further release the transport efficiency, it is necessary to upgrade the waterway and promote the organic connection of shipping system elements such as waterway, port and navigation aids, which requires increasing the application of artificial intelligence." Yan Xinping, an academician of the CAE Member and professor of Wuhan University of Technology, suggested that China is promoting the construction of major water transport projects such as the Pinglu Canal and the Three Gorges Water Transport New Channel, and should take this as a starting point to accelerate the transformation of the shipping system from a single breakthrough to a "ship cargo port, man machine environment" system. Look at the road again. Jingxiong Expressway is equipped with Baidu AI Cloud, which realizes "second level push" of important abnormal events through the cooperation of visual and language models; Didi Autonomous Driving and GAC Aion jointly create a new generation of pre installed autonomous vehicles... Autonomous driving collaborates with vehicle road coordination to enable more "smart" cars to drive on the "smart" road. The opinion insists on demand-oriented and scenario driven approaches, which means that through benchmark demonstrations, the entire industry can be driven to improve quality and efficiency. ”Wang Yunpeng proposed that taking the combination of assisted driving as an example, supporting key regional agglomeration innovation based on existing foundations, exploring provincial and ministerial linkage and vehicle road cloud integration collaboration, will promote the acceleration of commercial scale of assisted driving. Strengthening computing power, data, networks, and other aspects to ensure the deep integration of artificial intelligence and transportation cannot be achieved without the support of new infrastructure. Specific deployments will be made in terms of computing power, data, and network to strengthen element guarantees. In terms of computing power, it is suggested to strengthen the overall planning of industry computing power resources and rely on major transportation infrastructure to strengthen computing power support according to local conditions. According to the person in charge of Baidu AI Cloud, according to the needs of different scenarios, Baidu Intelligent Cloud can provide high-performance computing power cluster support for smart high-speed large models to ensure that event detection, emergency dispatching, etc. can obtain continuous and stable computing power. "In the future, we will deepen the application of large models in more road scenarios to serve the construction of smart high-speed." In terms of data, the next step will accelerate the construction of comprehensive transportation big data center, promote data sharing and high-quality data set construction, and better release the value of data elements. Data is the 'fuel' for the training and application of artificial intelligence big models, "said a person in charge of China Logistics Group. China Logistics has established a full chain data collection system of" person vehicle cargo yard road ", forming high-quality datasets for multiple typical scenarios such as transportation trajectories and multimodal transport. In the future, we will accelerate the construction of a digital product matrix to promote the development of the industry from 'experience based decision-making' to 'data intelligent decision-making'. In terms of the network, we will promote the integration and application of various network technologies, build an intelligent perception system and high-speed data transmission channels, and provide low latency, high reliability, and wide coverage connectivity capabilities for vehicle road collaboration, remote control, real-time monitoring, etc. (New Society)