Shulou(Shulou.com)11/24 Report--

The pace of market development is accelerating, and now the new energy vehicle revolution has gradually turned to the second half with intelligence as the main task.

On the development of the smart car industry, Li Ming, general manager of Jiangqi Group Co., Ltd., said that we are expected to catch up with and surpass Europe and the United States in the L4 stage. According to the relevant data, the American Society of Automotive Engineers (SAE) divides the autopilot technology into L0-L5 according to the degree of control of the vehicle control task, and the system mainly plays an auxiliary function at the L1-L3 level; when it reaches level L4, the vehicle driver will all be handed over to the system, while the difference between L4 and L5 lies in specific scenarios and full-scene applications.

There is no doubt that Chinese car companies need to make more efforts in technology research and development if they want to catch up with Europe and the United States in the L4 stage. But is the higher-level smart driving pursued by car companies what users really need?

01. From fuel vehicles to electric vehicles: the development of intelligent driving or overtaking cars in corners is a long process. After continuous innovation, research and improvement, cars have finally become what they are today with various forms and specifications. Then it is widely used in many fields of social and economic life. As one of the peaks in the field of civil science and technology, the threshold of automotive technology is not low, which takes a long time to accumulate and invest in patented technology.

In the era of traditional fuel cars, China's cars started late, and foreign brands have been studied for hundreds of years. After generations of wisdom precipitation, foreign brands have firmly occupied the top of the pyramid of engines, gearboxes and other technologies. For a long time, when it comes to foreign brands or joint venture brands, people's evaluation is "technology cow". Brands marked with foreign capital are often more likely to win the trust of consumers. Therefore, the development of the domestic automobile industry is shrouded in the shadow of foreign brands in the early stage, and it is even more difficult for domestic brands to make a breakthrough in traditional fuel vehicles.

Today, with the gradual rise of new energy vehicles and become the mainstream of development, global cars almost stand on the same starting line, and domestic car brands have taken the lead. Starting from the hybrid power, gradually entering the battery-based new energy era, electrification and intelligence are its two major characteristics. At present, the number of new energy vehicles in China ranks first in the world, and the momentum of development is fierce. Domestic major car companies want to achieve corner overtaking in the field of new energy vehicles, while China's new energy vehicles three electricity technology has made remarkable achievements, the pattern is stable, it is difficult to open the gap. As a result, car companies aim at intelligent driving in intelligence.

As a benchmark for smart driving, Tesla's market capitalization is among the most popular among the global mainframe factories. Its sales volume is less than 1/10 of Toyota's, but its market capitalization is three times that of Toyota. According to the research report of Guotai Junan Securities, there are many reasons for consumers to buy Tesla, in addition to the three electricity system, convenient charging experience, high value preservation rate, brand influence, its intelligent driving ability is one of the important considerations.

Tesla is the only mainframe factory that has developed the whole stack of autopilot and achieved mass production. After it was introduced into China, it gave full play to its catfish effect and aroused the competitive spirit of other domestic car companies. Car companies are in full swing to develop smart driving, and do not forget to mention their own smart driving when promoting new cars. Electric vehicles even more, it seems that there is no smart driving are embarrassed to compete in the huge car market.

Smart driving is becoming the next strategic highland in the eyes of car companies.

02. Medium and low-level intelligent driving has entered the development fast lane. At present, the penetration rate of L1 and L2-level auxiliary driving in China has increased rapidly. The active safety and driving function of L1 intelligent driving has gradually become standard, and has entered the period of mass popularization. In China, intelligent driving at L1 and L2 levels has been rolled out on a large scale, and Volkswagen-priced models speed up the L2-level program.

According to the China Automobile Daily, 4.766 million new cars with L2 smart driving were sold in China's passenger car market in 2021, accounting for 23.5%, an increase of 57.2% over the same period last year.

According to split data from Zoss Automotive Research, among the new cars sold in the first half of 2022, the L1 level has made the fastest progress in active safety, with a penetration rate of more than 30%. The driving function permeability of the L1 level has increased rapidly, from about 10% last year to about 20% in 2022. L2-level assisted driving is mainly concentrated in parking scenes, with automatic parking assisted APA penetration reaching 13.8%.

L3-L5 's intelligent driving is in the early stage of commercialization, and the self-research school represented by the new forces of car building and technology enterprises has made rapid progress and has begun to actively test the waters of the urban road scene. NOA pilot driving assist is a high-level driving assistance feature that has been mass-produced at present, and only new head-building forces such as Wei Xiaoli and head technology companies such as Huawei have achieved high-speed NOA.

The NOA of urban roads is the next step to conquer. As the urban road scene in China is complex and changeable, the research and development of urban NOA is much more difficult than that of high-speed NOA. Compared with high-speed NGP, the code amount of urban NGP increases to 6 times; the number of perceptual models increases to 4 times; the part related to vehicle behavior includes prediction, planning, control, and related code amount increases to 88 times.

At present, Xiaopeng and Jihu have respectively begun to push the pilot of urban NOA. However, Xiaopeng is only conducting small-scale tests on some specific roads in Guangzhou, and Jihu, which is equipped with Huawei technology, only makes a closed push to Beta users in Shenzhen.

Of course, the research and development of the foundation and technology of China's intelligent automobile industry is still relatively weak. Li Ming, general manager of Jiangqi Group Co., Ltd., mentioned that greater efforts are needed in vehicle research and development, sensors, and so on. especially in the control decision-making and implementation technology, there are still some gaps with the international advanced level, and the shortcomings of automotive operating systems and chips are still obvious. Secondly, the new intelligent manufacturing capacity to adapt to intelligent car manufacturing needs to be further improved, and the application demonstration of intelligent cloud management and control needs to be further unified.

Finally, the data collected at present is not enough to support L4 self-driving vehicles to make independent decisions, and the cost of vehicle-road-cloud integrated infrastructure is still relatively high.

It can be seen that the development of intelligent driving in our country is accelerating, and it still takes time to realize overtaking in corners.

03. In smart driving, no matter it is the new force of car building or the traditional mainframe factory, they have made great efforts in intelligent driving and launched a fierce competition. Mainstream car companies are scheduled to achieve L4 in 2024 or 2025, technologically take the sensor fusion route, hardware pre-embedding represented by lidar has begun, and the construction of data closed loop is the key capability.

Hardware pre-embedding is a new mode produced by car companies driven by autopilot, that is, through hardware pre-embedding in the car, through continuous software upgrades, users can get a better experience. With the narrowing of the hardware gap between independent and joint venture models, consumers who are mainly physical perception are gradually turning to in-car experience perception. Hardware pre-embedding is an important mode to improve user experience.

Hardware embedded arms race has begun, how to make good use of data and build data closed loop is the key. Among the models listed in 2022, such as Weilai ET5, ideal L9, Xiaopeng G9, Nezha S, Guangzhou Automobile E'an AION LX Plus, all of them are equipped with more than one lidar, as well as a high-precision map positioning unit, and the hardware pre-buries of various manufacturers are already in place, implicitly indicating that these independent Chinese players will take the fusion route of sensors and high-precision maps.

It is agreed that the future driving mode must be autopilot, but the current driving technology is still difficult to achieve. Even Apple has to succumb to L5 self-driving technology. Not long ago, Apple announced that it would abandon its plan to launch a "L5" fully self-driving car. Compared with Apple, which is "ambitious", domestic enterprises seem to be at ease and handy.

The earliest intelligent driving company in China needs hundreds of degrees, and launched the unmanned vehicle project in 2013. Baidu's advantage lies in AI technology reserves and high-precision maps, L4-level self-driving data, Baidu has accumulated more than 10 million kilometers, and the simulation data is as high as 1 billion kilometers. The housekeeper's Apollo intelligent driving system has also been iterated to version 6.0, with more than 70 car company partners.

However, the cost of resounding Baidu smart driving has been high, which has become a fatal disadvantage of Baidu smart driving. Baidu's fifth-generation Robotaxi costs as much as 480000 yuan, but there is not enough market to share the high cost, so the scale effect can not be realized.

Huawei, another technology company, has previously claimed that it does not build cars, but its voice is not small in the new energy car circle. In intelligent driving, Huawei follows the technical route of multi-sensor fusion. The Huawei intelligent car solution carried by BAIC Jihu, which is equipped with 3 lidars, 6 millimeter wave radars, 12 cameras and 13 ultrasonic radars, as well as Huawei chips with computing power up to 352Tops, has reached self-driving applications above L3 and can achieve 360-degree full-angle perception. Can realize the autopilot system to structure the road into the urban road.

Although the importance of Huawei's auto business to the development of smart cars in China is self-evident, Huawei still faces several risks: first, high-end chips cannot be manufactured, and inventory problems need to be solved; second, whether Huawei's car companies can provide a large amount of data for Huawei's self-driving system iteratively; third, how Huawei can win the trust of its partners. Only by solving these problems can Huawei's smart driving business go a step further.

Intelligent driving undoubtedly occupies a core position in the development of intelligent cars, and in the eyes of automobile companies, it is tantamount to "getting smart drivers to get the world". To be sure, consumers pursue intelligent cars that are more technological, more comfortable and more convenient, and how to reasonably match intelligent driving to user application and experience, it will be an unavoidable test question for enterprises to develop and study intelligent driving.

This article comes from the official account of Wechat: automotive News Agency (ID:automobile-news), by Zeng Qian

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.