On March 17, Pony.ai and Baidu announced that they have obtained the first batch of "unmanned off-vehicle long-distance stage" (fully unmanned) demonstration application licenses in Beijing High-level Autonomous Driving Demonstration Zone, and have been approved to start the fully unmanned autonomous driving travel (Robotaxi) service in Beijing Yizhuang.

 

It is understood that the demonstration application of "unmanned vehicle remote stage" (unmanned vehicle in the whole vehicle) must rely on the integration of vehicle and road, and use the cloud control platform to ensure the safety of autonomous vehicles by aggregating more dimensions, more refinement and more real-time data on the vehicle, roadside and third-party platforms. "If L4 autonomous vehicles want to be truly applied, they must rely on the cooperation of the road and the cloud." Dai Yifan, assistant dean of Suzhou Automotive Research Institute of Tsinghua University, expressed his views on the construction of vehicle-road-cloud integration in an interview with China Automotive News recently.

 

Vehicle road cloud breaks down information barriers

 

In October 2021, the Beijing Intelligent Connected Vehicle Policy Pilot Zone opened the autonomous driving unmanned road test for the first time, clearly dividing the "unmanned" test into three stages: "co-driver manned", "unmanned in the front seat, manned in the rear", and "remote outside the vehicle".

 

Therefore, due to the actual needs of the above three stages, it is important to establish a good data closed loop to promote the development of vehicle-road-cloud integration. "At present, the data collected by autonomous vehicles is feasible for technical iteration and information service early warning, but to support the decision-making and control of autonomous vehicles, the existing data quality needs to continue to be cultivated." Dai Yifan believes that excluding technical factors, the high cost is undoubtedly the core that restricts the rapid formation of effective collaboration in the vehicle-road cloud. He said that judging from the bidding information of vehicle-road cloud collaborative projects published by some cities, a single project can be as little as 10 million yuan or more than 100 million yuan. How to recover the cost or even profit in the end of such a huge project, there is currently no very mature case in the industry.

 

From this level, the cost dilemma has become a resistance for some autonomous driving companies to continue to take the integration of "vehicle, road and cloud". "This requires a high degree of awareness, comprehensively promoting the construction of vehicle-road-cloud collaborative infrastructure, and providing support for autonomous driving testing and demonstration." Dai Yifan said that the cloud control platform involved involves the infrastructure construction of vehicle-road-cloud collaboration, and the cloud control platform means that the opening of these data, standards, and protocols can help the good operation of relevant demonstrations.

 

Commercial exploration needs to be accelerated

 

The vehicle-road-cloud integration system integrates the physical space and information space of people, vehicles, roads and clouds through a new generation of information and communication technology, and realizes the safe, energy-saving, comfortable and efficient operation of the intelligent networked vehicle transportation system based on the collaborative perception, decision-making and control of the system, and has the advantages of standardized access capability, common basic capability, open sharing capability, intensive efficiency enhancement capability and mode replication capability.

 

In Dai Yifan's view, in the construction of intelligent network demonstration area, to solve the integration of vehicle and cloud, as well as the coordination between intelligent network demonstration areas, it is necessary to open up information between demonstration areas in various places, because from automatic driving testing, intelligent network demonstration, and future commercial applications, coordinated development is required.

 

In this regard, Dai Yifan believes that in order to achieve the above goals, a series of work needs to be done.

 

First of all, to establish and improve relevant laws and regulations, of course, legislation at the national level takes time, and the competent authorities also encourage local intelligent network demonstration to be piloted first, although the legislation at the local level is only regional, but can be based on reality in the region, continue to try, gain experience, learn from each other, and gradually improve, so as to further provide a basis for national legislation.

 

Secondly, with the increasing number of application demonstrations of intelligent network and autonomous driving in many places in China, relevant commercial operation demonstrations have also appeared, but in fact, they have not achieved commercialization in the true sense, that is, their business models have not been fully established, for example, how to charge, charging standards, etc. are not consistent, and the current fees are not enough to offset the cost of vehicle operation and infrastructure construction. In short, its commercial closed loop has not yet been fully established.

 

"Therefore, the breakthrough in the commercialization of autonomous driving is not only related to infrastructure construction, but also related to the continuous improvement of relevant technologies, standards, regulations and systems." Dai Yifan said that the continuous maturity of autonomous driving technology can make the product continue to mature, and after the product is stable, it can save operating costs, bring about the continuous reduction of the price of the product itself, and lead to the continuous reduction of the cost of operation, so that the final commercial closed loop of autonomous driving will be formed. "This process is a long way off, and the commercialization of autonomous driving still has a long way to go." Dai Yifan believes that the core is to improve the technical level, improve the maturity of the product, and reduce the cost of the product, which are the foundation of the foundation that needs to be consolidated.

 

Originally published on China Automotive News on March 20, 2023