“We have also noticed that the current autonomous driving research and development data is no longer simply winning by quantity, but there is a greater demand for high-quality data such as extreme scenario operating conditions.” At the 2026 World Intelligent Connected Vehicle Conference media roundtable on July 27, Ge Yuming, director of the Vehicle Networking and Smart Transportation Research Department of the China Academy of Information and Communication Technology, said in response to questions from reporters and other media. From an enterprise perspective. The first is to accelerate the evolution of the supply model towards diversified collaboration. On the basis of actual vehicle acquisition models that rely on data collection vehicles or mass production data return, further develop the advantages of all-weather and multi-view deployment of roadside sensing and communication infrastructure, and explore the use of more real complex traffic scene data that can be obtained by roadside sensing for autonomous driving research and development. The second is to strengthen the development and application of generative autonomous driving data production technology. Accelerate research and development of data generation technology such as exploration world models, and support data generalization for long-tail scenarios such as bad weather, special road sections, complex traffic games, and dangerous behavior. The third is to explore the flow of trustworthy and controllable data across subjects. Explore high-value scenario data circulation across entities through data circulation platform facilities such as trusted data spaces in the automotive industry.

“We have also noticed that the current autonomous driving research and development data is no longer simply winning by quantity, but there is a greater demand for high-quality data such as extreme scenario operating conditions.” At the 2026 World Intelligent Connected Vehicle Conference media roundtable on July 27, Ge Yuming, director of the Vehicle Networking and Smart Transportation Research Department of the China Academy of Information and Communication Technology, said in response to questions from reporters and other media. From an enterprise perspective. The first is to accelerate the evolution of the supply model towards diversified collaboration. On the basis of actual vehicle acquisition models that rely on data collection vehicles or mass production data return, further develop the advantages of all-weather and multi-view deployment of roadside sensing and communication infrastructure, and explore the use of more real complex traffic scene data that can be obtained by roadside sensing for autonomous driving research and development. The second is to strengthen the development and application of generative autonomous driving data production technology. Accelerate research and development of data generation technology such as exploration world models, and support data generalization for long-tail scenarios such as bad weather, special road sections, complex traffic games, and dangerous behavior. The third is to explore the flow of trustworthy and controllable data across subjects. Explore high-value scenario data circulation across entities through data circulation platform facilities such as trusted data spaces in the automotive industry.