ICT, road test, software and smart car are three short boards to be supplemented
smart car shows us a beautiful blueprint. When L4 and L5 automatic driving is popularized, our hands will be liberated from the steering wheel and we can really enjoy sitting in the car and looking at the scenery outside the window. However, the realization of this process still faces many challenges. At the 2020 World Smart car conference held recently, industry experts and scholars deeply discussed and analyzed the weaknesses of China's smart car industry in the three fields of ICT (information and communication technology), road testing and basic software. They believed that only by consolidating these three foundations can China's smart car achieve leapfrog development
in the final analysis, the smart car should realize the communication connection between vehicles, between vehicles and people, and between vehicles and. Liuyunjie, academician of the Chinese Academy of engineering, said: "ICT is the foundation of smart cars. The development of high-level smart cars puts forward new requirements for ICT."
for the smart car, the inside of the car can be regarded as an internal space, and the communication in the car becomes the foundation of single car intelligence. The sensing system transmits the sensor flow to the corresponding computing unit, the decision-making system allows multiple computing units to process the sensor data together, and the control system transmits the data to the actuator... Finally, the laser radar can realize the sensing function of automatic driving after a series of processing. Liuyunjie said, "the launch cycle of the lidar is required to reach 150ms. If the delay and jitter of the communication in the vehicle are greater than 150ms, the command may be wrong."
communication outside the train is the basis of communication, and there are certain requirements for delay. The delay demand of L2 level automatic driving is 20 ~ 100ms, L3 is 10 ~ 20ms, and the delay of L4 and L5 automatic driving is less than 10ms, or even less than 1ms. " Today's Internet technology is far from reaching this standard. " Liuyunjie said
what kind of network can meet the requirements of smart car driving? Liuyunjie said: "the end-to-end delay of traditional IP forwarding is unpredictable due to the randomness of queuing, and may cause tens of milliseconds of jitter or even packet loss; while the end-to-end delay of dip forwarding is deterministic and controllable, and the delay jitter can be controlled at or below the level of 100 microseconds."
in addition to certainty, there must be certainty to match. Major foreign automotive standards organizations and industry alliances have been actively exploring the application of TSN (time sensitive network) in the automotive network. The on-board TSN architecture mainly includes three parts: regional group architecture, domain controller architecture and central customs architecture. TSN architecture must be integrated with 5g technology to play a role. If the remote control of automatic parking scenario requires the network end-to-end delay to be as low as 5ms, TSN fusion 5g technology can effectively reduce the delay
liuyunjie said: "in the future, the deterministic network technology will help the development of smart cars. The end-to-end deterministic network based on TSN, TSN ＋ 5g, dip ＋ Geni and other technologies will greatly help the technical development of the internal and external networks of smart cars."
one of the characteristics of artificial intelligence is to accumulate data through continuous road operation. It provides reference for the decision-making of intelligent control system. By the end of September 2020, nearly 30 provinces and cities in China have issued detailed rules for the implementation of road testing, with more than 2800 kilometers of open road testing mileage, and more than 70 enterprises have obtained more than 430 road testing licenses in total. Among them, 87 licenses were issued in Beijing, 119 licenses were issued in Shanghai, and 53 licenses were issued on the couch
it is understood that the test management of smart car follows the way of central planning and local self-management. Relying on the test management specifications of the three ministries and commissions, local governments have formulated localized test management rules. Therefore, there are differences in specific operations, and there are obstacles in the interconnection of test results, test licenses and test data
according to public data, in 2019, Shanghai, Jiangsu, Anhui and Zhejiang signed a mutual recognition cooperation agreement on road testing; Beijing, Tianjin and Hebei have signed a mutual recognition agreement on road testing between Beijing, Tianjin and Hebei; Hebei, Shandong, Guangzhou, Shenzhen, Wuhan and other regions recognize the closed test reports or test licenses of other provinces and cities. Gong Weijie, Secretary General of China intelligent Alliance for automotive industry innovation, said: "there are still obstacles to mutual recognition of road tests in China. Only a few cities recognize each other, and some mutual recognition remains at the agreement stage, but has not been implemented."
there are many factors affecting road test interaction, including vehicles, yards, people, and management. In terms of test types, there are great differences among different regions. For example, Beijing is divided into road test, formation test, manned test and cargo test
in order to promote mutual recognition and interoperability, the innovation alliance, in conjunction with experts and scholars of automobile enterprises, has developed the road map for promoting the interaction of intelligent United Automobile Testing (hereinafter referred to as the road map). The goal is to realize mutual recognition of test results for common basic test items, local special test items and the same test items, and reduce the test content applied in different places. The test vehicle can exchange the test license in other places through the test notice, or supplement a small amount of data to exchange the test license in other places, which is the initial sign notice for the engineering utilization of modern hydraulic technology, reducing the approval process of application data and improving the efficiency of license application. The ultimate goal is to carry out remote testing without changing license plates, completely eliminate the regional threshold and achieve cross regional continuity testing
the roadmap also establishes a detailed schedule, unifies all test projects in 2021, and optimizes the service capacity of the test area; From 2010 to 2022, unified test driver approval requirements, established test driver management mechanism, and established test supervision data transmission standards; After 2022, it is necessary to achieve mutual recognition of test notices, unify the requirements for vehicle test types, unify the classification standards for test roads, unify the requirements for submission of test application materials, and unify the requirements for preparation of test notices
the hardware equipment of the smart car is very important, as is the software. However, in the past, China has not compiled a complete research report on automotive basic software, so the decision-making of government departments may be biased. A few days ago, caobin, the first rotating chairman of the China Automotive basic software ecological Committee, on behalf of the alliance, released the white paper on the development of China's automotive basic software 1.0, which became the first research report with reference and guiding significance
the problems related to China's automotive basic software are mainly summarized into three aspects: first, technical problems; Second, application; Third, standardization. Technical issues cover a wide range. In the era of smart cars, the automotive architecture has undergone earth shaking changes. From the past distributed architecture to the domain controller based architecture. In the future, it will develop towards the architecture based on the central computer. This change involves virtual technology, operating system, functional security, computing support, etc. each technology has its own internal core and key, which requires the industry to work together to complete the development of automotive basic software
in terms of application, automotive basic software solutions are generally faced with problems such as high development cost, high project management cost, high capability requirements, incomplete industrial chain and R & D system, insufficient technical service support and insufficient reserve of professionals. In terms of standardization, there are also problems. For example, many enterprises participate in the formulation of basic software standards, but the standards involve multiple core interests, and it is not easy to reach a consensus in a short time
Cao Bin believes that before the birth of basic automotive software, automotive function development was realized through "independent software + specific hardware", and there was no unified interface between software and hardware. When the same function matched different models, the software could not be reused because of hardware differences
under the trend of "new four modernizations" in the global automotive industry, intelligence has become an irreversible trend in the development of the automotive industry. The era of software defining the automobile has come. In the future, the definition of the automobile is no longer the traditional technical performance indicators, but the software empowerment with artificial intelligence and human (1) speed reduction while improving the output torque and machine interaction as the core
automotive basic software is a new field. Domestic automotive enterprises are in the same lineup as those in developed countries and regions in Europe and the United States, and some are even at the leading level. For example, in the United States, Google, waymo and other companies are actively developing smart cars. In China, Alibaba, Weilai, Chang'an Automobile and other companies are also strengthening the research and development of smart car technology, which has considerable strength
in the future, cars will change from simple means of transportation to smart terminal products that connect people's work and life, energy and smart cities. The barrier of traditional automobile industry is relatively high, but the basic software of automobile is different. Intelligent connection technology is a new thing for us, and the same is true for European and American automobile enterprises. The new automobile architecture is still in the process of forming, and the key is to seize the opportunity. China is the largest auto production and marketing region in the world. European and American auto enterprises have increased their R & D investment in China. The Chinese auto market provides a market foundation for the local industrial chain to catch up
recently, the central government proposed a new pattern of promoting domestic and international dual cycle development, creating a good policy environment for domestic automobile enterprises to seize the opportunity. Internationally, as the situation of epidemic prevention and control is not optimistic, European and American automobile enterprises have begun to pay attention to Chinese software suppliers to avoid affecting the freezing effect. Some European and American R & D personnel cannot come to China. However, many products to be listed in the next year need a lot of software development, which provides a good opportunity for Chinese suppliers and is expected to cultivate a group of competent technical talents through market practice
China has taken the lead in 5g field, and is not backward in road testing. The shortcomings of automotive software are being accelerated to make up. From a comprehensive point of view, it is just around the corner for China to take the lead in realizing the revolutionary breakthrough of smart cars. (wanrenmei)Overview of environmental protection shopping bags becoming popular