news

Tencent News "First Line"

Author: Lin Keying Editor: Liu Peng

The automotive industry has entered the second half of the intelligentization stage, and the debate over the route of autonomous driving has not stopped. Among them, single-vehicle intelligence and vehicle-road collaboration are one of the focal points of the controversy. On August 17, Li Keqiang, an academician of the Chinese Academy of Engineering, said at the Fourth Shenyang Intelligent Connected Vehicle Conference that vehicle-road-cloud integration is not contradictory to single-vehicle intelligence, but an upgraded version of single-vehicle intelligence. Once single-vehicle intelligence is done well, vehicle-cloud collaboration will naturally derive, and there is no question of who denies whom.

byTeslaTaking the new generation of FSD as an example, Li Keqiang believes that its "shadow model + end-to-end" is actually an application model of vehicle-cloud collaboration, rather than traditional single-vehicle intelligence, and vehicle-road-cloud integration is a comprehensive upgrade of single-vehicle intelligence + shadow model. "The existing single-vehicle intelligence can only approach the level of excellent drivers, which limits the upper limit of the level. The introduction of roadside perception and computing enhances the vehicle-road-cloud collaboration model, which has the potential to surpass human driving level and raise the upper limit."

Li Keqiang said that there are still two major problems in the technical closed loop of vehicle-road-cloud integration. First, the use of vehicles and system operation are mainly based on single-vehicle intelligence, and the acquisition of data for vehicle-road-cloud integration is very passive, and few car companies actively or voluntarily connect data to the platform; second, the current vehicle-cloud collaborative integration is often a chimney-shaped architecture, with only intelligent driving data, and no cross-domain sharing.

"Intelligent connected vehicles are high-tech products, and the industrialization of high-tech should have a technical closed loop and a commercial closed loop. A technical closed loop may not necessarily mean a commercial closed loop, but without a technical closed loop there will definitely be no commercial closed loop." Li Keqiang expressed the hope that a technical closed loop can be formed as soon as possible through application pilots.

After achieving a closed technology loop, how can we form a closed business loop? Li Keqiang stressed that we should focus on implementing application channels and focusing on pain point scenarios. He summarized eight application scenarios, including smart transportation, smart sanitation, road logistics, smart passenger car companies, data value-added services, urban logistics, smart charging and discharging, and smart mines. "If we really want large-scale infrastructure construction and cannot form a closed business loop, there will be problems."