[Autohome? Industry] Since the beginning of this year, the application of autonomous driving technology in China has continued to advance. Currently, cities including Beijing, Shanghai, Chongqing, Guangzhou, Changsha, Shenzhen and other cities are actively conducting manned research and development tests of Robotaxi (self-driving taxis) on urban roads. To some extent, this is a race to gain entry into the future mobility market. In this race for the future, no one wants to lose. Baidu's driverless cars are already available for use in Changsha, Beijing, Cangzhou and other places; AutoX and Didi have successively launched driverless taxi services in Jiading, Shanghai; Pony.ai and WeRide have also launched similar driverless taxi services in Guangzhou Serve. Recently, Zhongzhixing's self-driving vehicles have been undergoing public road testing in Fengxian District, Shanghai. Is another player coming to carve up the Robotaxi market share? Zhongzhixing is a 2-year-old start-up autonomous driving company. Wang Jin, the chairman and CEO of Zhongzhixing, was the former senior vice president of Baidu and the founder of the autonomous driving division. Compared with the leading echelon of autonomous driving, for China Intelligent Technology, which entered the market relatively late, what confidence does it have to come from behind? On November 10, after experiencing Zhongxing’s autonomous vehicles at close range and having in-depth conversations with senior management, we tried to find an answer.

■ Positioned as Robotaxi, the Lincoln MKZ is located in the Lingang Nanqiao Science and Technology Oasis in Shanghai, and has attracted a number of start-ups. Zhongzhixing is one of them, where autonomous driving technology is being repeatedly tried and tested. Entering Lingang Nanqiao Science and Technology Oasis, two cars with "large equipment" on their heads have been waiting for visitors on the roadside for a long time. This is a Lincoln MKZ model. In order to support L4 level autonomous driving, it has added very large and complex sensors and a complex set of computing and processing units, and applied 5GAI technology to it. What specific configuration plan is it? Chen Yinzi, vice president of technology of Zhongzhixing, introduced to Autohome that at the perception level, that is, the "eyes" of driverless driving, the Lincoln MKZ test car integrates 2 monocular cameras and 4 lidars (overhead, front and rear). Each has 32 lines at the end, and the other is only for data collection), 6 millimeter wave radars, satellite navigation and inertial gyroscopes and other hardware systems. (PS: The reason why it is all exposed is for the convenience of viewing. When it is put into mass production later, the beautiful design will be considered and customized.) Inside a large box in the middle is the driverless algorithm unit, which is equivalent to driverless driving. The brain, engineers and programmers monitor the vehicle through remote operation. It is reported that the maps are independently collected and produced by Zhongzhixing. They can collect information around the vehicle in real time and generate a three-dimensional map with an accuracy of currently up to 2cm. During the test ride experience, the map information displayed on the central control screen can clearly prompt the surrounding vehicle conditions, as well as the waiting time for traffic lights, etc. During the entire test ride experience session, due to the relatively simple road conditions, the Lincoln MKZ test car completed driving in various scenarios such as going straight, changing lanes, turning left at traffic lights, passing traffic lights, and avoiding pedestrians, without the intervention of a safety officer throughout the process. . If we talk about flaws, during the test ride, we also noticed problems such as sudden obstacles on the roadside, which caused sudden braking, a slight force when changing lanes, and difficulty in ensuring smoothness when turning left without protection. However, these problems also appeared in other Robotaxi users. In experience. According to reports, Zhongzhixing provides two driverless experience modes. One is the sports mode, which involves more intense speed increases such as starting and turning during driving; the other is the comfort mode, where the vehicle speed is slower and the execution process is gentler. From the perspective of technical level, Zhongzhixing is not inferior to several leading domestic autonomous driving companies in terms of vehicle control, planning and decision-making, positioning, and integration. Zhongzhixing's positioning is also very clear. What it wants to do is to provide the most enjoyable travel services based on driverless driving. At the 2019 Boao Forum for Asia, the Hongqi H7 PHEV driverless car jointly developed by China Intelligent Technology and China FAW made its debut, providing test ride experience services to visitors.

■Vehicle-road collaboration, opportunities for overtaking in corners. Even if the level of intelligent bicycle technology is similar, for Zhongzhixing, which has not yet been put into operation, it is still a "metaphysics" to surpass the top few companies in China. ”, not to mention Waymo, which has already gone to the extent of canceling safety officers in the United States. Looking at a set of data, the United States ranked self-driving companies around the world in March 2020 (pictured below: Navigant research report). There are more than 20 self-driving companies, and only one Chinese company can be on the list. In view of this, Wang Jin believes that vehicle-road collaboration is an opportunity for Chinese companies to overtake in corners. "It will take at least 2-3 years for China to catch up with the top three in the United States in terms of bicycle intelligence. And even if we are given two years, we may not be able to complete the leap of eliminating safety officers, mainly because of China's Traffic conditions are much more complicated than in the United States," Wang Jin explained. This is why China Intelligent Technology shifted its focus from bicycle intelligence to vehicle-road collaboration in 2019. When CIIC was established, it was on the eve of 5G commercialization. Therefore, in the architectural design of the entire autonomous driving system, Zhongzhixing has already targeted the application of 5G technology. "Interfaces have been reserved for the integration of 5G and AI, involving modules such as perception and positioning, to avoid later tinkering." said Chen Yinzi, vice president of technology at Zhongzhixing.

Compared with other companies, China Intelligent Bank has also been "stuck" in time to a certain extent. “In terms of safety, bicycle intelligence cannot predict in advance in many scenarios, and there are blind spots in perception, such as ghost probes, blocking obstacles, etc.” Chen Yinzi explained, “Vehicle-road collaboration can achieve beyond-line-of-sight perception. . The roadside unit accurately transmits road information that cannot be sensed by the vehicle to the car through its own sensors, eliminating many detection blind spots. "According to Zhongzhixing's understanding and definition, the key to vehicle-road collaboration lies not in the hardware of the roadside unit, but in it. How much smart software goes into the hardware of these roadside units. It is not just a sensor. It needs to identify whether it is a person or a variety of vehicles. It also needs to predict the trajectory and behavior of the vehicle. For example, will the vehicle slow down at the intersection and obey the traffic rules? Will the vehicle run a red light? From the perspective of artificial intelligence, these predictions can be made in advance, and the accuracy can reach or exceed 99%. It is reported that Zhongzhixing has implemented basic vehicle-road coordination functions in Shanghai Lingang and Shanghai Fengxian. "In the future, we will overlook our transportation network in front of the central command system, where sensors and roadside units are densely deployed to command vehicles and other traffic equipment flowing on the road." Wang Jin said.

■A bargain, accelerating industrial integration. Of course, no company can go it alone on the road-vehicle collaboration route. It needs government support and the integration of the entire industry chain. In fact, China's good policies are also providing support for China Intelligent Technology's route: In February 2020, 11 ministries and commissions including the National Development and Reform Commission jointly issued the "Smart Vehicle Innovation and Development Strategy" pointing out that vehicle-road collaboration is an important development direction in the future. With the support of 5G and new infrastructure, China is most likely to achieve vehicle-road collaboration. Wang Jin pointed out that vehicle-road collaboration requires the support of 5G communications and the intelligent transformation of roads, including connecting wires, networks and many sensors on the roads, which also requires strong government support and promotion. Regarding the cost of vehicle-road collaboration and bicycle intelligence, which one is more expensive? China Intelligent Technology Co., Ltd. has also calculated that vehicle-road collaboration has cost advantages. Smart roads can reduce the smart cost of each bicycle by 20,000 yuan. In the first half of 2020, the number of cars in the country was 270 million, which can save a total of 5.4 trillion yuan. As of 2019, the total mileage of China's highways reached 4.8465 million kilometers, and the cost of vehicle-road coordinated reconstruction for a single kilometer is about 800,000-1 million yuan. Roughly calculated, the entire vehicle-road coordinated reconstruction cost will be 3.8 trillion-4.8 trillion yuan, which is relatively Smart bikes are cheaper. In terms of industrial integration, China Intelligent Bank is also actively seeking cooperation. In October 2020, China Automotive Chuangzhi and China Intelligent Technology Co., Ltd. reached a strategic cooperation. The two parties will jointly explore 5G smart vehicle-road collaboration through multi-field business integration and achieve common development; in September of the same year, the conference was held in Shanghai At the 2020 World Artificial Intelligence Conference, China Telecom Technology Innovation Center, Shanghai Municipal Commission of Economy and Information Technology, and China Intellectual Property Office officially signed a tripartite cooperation agreement on "Key Technology Research and Industrial Applications of Intelligent Vehicle-Road Collaboration" to develop the direction of vehicle-road collaboration in intelligent transportation. Carry out cooperation to achieve large-scale verification of smart vehicle-road collaboration, and take the lead in launching commercial-level promotion and demonstration operations in Shanghai.

■Summary: Previously, China Intelligent Technology’s plan was to launch 100 self-driving vehicles for operation on the road in 2020. However, according to on-site understanding, China Intelligent Technology’s plan has changed. It will not launch vehicles for road operation at the moment, but Vehicle-road coordination will be completed within the next 2-3 years. Many experts in the industry have judged that it will take at least five years for vehicle-road collaboration to be rolled out on a large scale. If the opening up of L4 autonomous driving laws and regulations progresses slowly, this time will be even longer. This means that although Robotaxi is backed by a trillion-yuan travel market, for companies, only continued investment can support this expensive experiment. Whether Robotaxi can survive until the dawn of time and achieve real profitability, China Intelligence Bank still has a long way to go if it wants to be the winner.

(Text/Autohome?Peng Fei)