It is an indisputable fact that the growth of smart phones has slowed down. Without this momentum, where will the next economic growth momentum come from? Potential industries such as 5G, new energy vehicles, autonomous driving, and the Internet of Things are all eagerly taking over, but the problem is that it took ten years for smartphones to appear to be saturated, and the decline started from the peak, and these markets are climbing upwards. It's still in the initial stage, who will be able to win in the end? Perhaps the first step is to see which industry can create a trillion-yuan market like mobile phones.
When will the scale reach 100 billion yuan?
With the help of global players, autonomous driving has become the focus of industry chain manufacturers. The governments of China, the United States, Japan and Europe have been constantly removing obstacles to their industrial development.
Therefore, the industry optimistically predicts that 2020 will be the year of commercialization of the autonomous driving market. From 2020 to 2025, the annual compound growth rate of self-driving passenger car sales will exceed 80%; together with the sales of self-driving vehicles and the self-driving travel service market, the total domestic market will reach approximately 40 billion in 2020 and 2030. US dollars and 550 billion US dollars.
The numbers are very eye-catching, and in terms of the popularity of autonomous driving, the key factors are performance and price. From a price point of view, perhaps expensive lidars, solid-state radars, etc. will drop to an acceptable level in the next five years, but the decisive factor is whether autonomous driving can reach the maturity of the technology in the foreseeable short term. And most importantly, supporting regulations and infrastructure can "evolve" together to become an assist.
It seems not too far away to "take the baton" as a growth driver, and in the new wave of technological change in a century-old car driven by autonomous driving, completing the upgrade or shuffling should be the best choice for participating players.
Will L3 be realized in 2020?
For the level of autonomous driving, there are currently a large number of mature mass production cases of L1 and L2 ADAS. After the Audi A8 and other foreign car parks released L3 (conditional autonomous driving) models, domestic automakers such as Chery, Changan, Great Wall, Xiaopeng, and Internet giants such as Baidu, Tencent, and JD.com have successively implemented L3 and even L4 (high automation) models. ) Technology to arrange troops.
Some analysts say that L3 autonomous driving has entered the test vehicle stage, and it is expected to reach the mass production level in 2020 and is expected to eventually achieve L3 level by 2025. Relevant representatives of Weilai pointed out that at present, autonomous driving is still in the critical stage of transition from L2 to L3. However, with the maturity of technologies such as chips, algorithms, and high-precision maps, and the continuous improvement of policies and regulations, it is expected that starting from 2020, L3 The autonomous driving market will usher in an explosion.
But there are also views that tend to be cautious. "From the current point of view, L3 autonomous driving technology still has a long way to go, which involves the influence of many factors from technology, industry chain, laws and regulations to market recognition." STMicroelectronics (ST) Asia Pacific Automotive Business Promotion Senior Manager Sun Guobin believes that “at the moment, ADAS+ solutions that fit L2 autonomous vehicles will be deployed on a large scale.â€
During the evolution process, some autopilot companies choose to enter L4 directly, while others start from L1 and carry out the iterative upgrade of L2-L3-L4. Sun Guobin is optimistic about these two methods. Sun Guobin explained that in order to solve application problems such as no robotaxis (a smart driverless taxi), it is necessary to achieve the highest level of autonomous driving. For most passenger cars, ADAS is sufficient, but this market is not large enough, and the target applications are very similar, so the two paths will overlap and cross. A relatively small number of hardware and algorithms used in autonomous vehicles will be stripped out and used to solve the mass market ADAS problem.
Multi-technology integration into an inevitable way?
In the sensor solutions for autonomous driving, vision, laser, and millimeter-wave radars are all responding to technical requirements in an attempt to win a C position.
Sun Guobin believes that all sensing technologies have advantages and disadvantages, but it is clear that computer vision technology has the potential to independently solve this level of autonomous driving problems. Companies such as Mobileye have now developed prototype vehicles equipped with vision technology and demonstrated the feasibility of computer vision autonomous driving on highways. The next step will prove its feasibility in urban road conditions.
"Sensor fusion is the general direction. Cost considerations and technological breakthroughs are the key. At the same time, achieving this goal at a reasonable cost is the ultimate goal." Jerry Cui, Automotive Marketing Manager, Greater China, ADI Automotive Electronics Business Unit, is optimistic about the sensor fusion solution. .
However, Sun Guobin also emphasized that in order to achieve a higher level of autonomous driving and make the system more robust, it is necessary to use other sensors with vision technology, such as radar, lidar, ultrasonic, and far-infrared sensors. In addition, V2X communication and precise positioning technology are needed to improve the performance of autonomous driving. The integration of multiple technologies is an inevitable way to realize intelligent driving. In 2020, vision and radar assisted driving systems will become standard equipment for most vehicles.
Therefore, the brain of autonomous driving, the computing chip, has naturally become the top priority, and the old and new forces are also in intensive confrontation. Nvidia, Intel, Qualcomm, Horizon, Shenjian (acquired by FPGA giant Xilinx) and traditional automotive semiconductor manufacturers NXP, Renesas, Infineon, and TI are all competing.
This also increases the requirements for computing platforms. "The requirements for the accuracy, detection range, and robustness of the sensor to the weather and the surrounding environment are getting higher and higher, and the requirements for the computing power and real-time performance of the computing platform are also getting higher and higher." Sun Guobin said, "These The system currently focuses on mission-critical tasks. Functional safety is the initial goal, and the ability to achieve safe parking of vehicles is the minimum requirement. Some OEMs have higher requirements and adhere to the concept of complete system redundancy to enhance fault tolerance."
Nvidia is gaining momentum in this field, and a number of domestic startups are also trying to share the pie, such as Heico Motors and Huanyu Zhixing.
At the software level, the perception, decision-making, and simulation algorithms for autonomous driving software also need to be continuously advanced. It is worth noting that technologies such as voice interaction, gesture interaction, HUD, Internet of Vehicles services, and Gigabit Ethernet are also popular on the autonomous driving track and have a lot to do.
Is the critical point for large-scale applications coming?
It can be said that autonomous driving technology has reached the critical point of large-scale mass production applications. In addition to the continuous improvement of technology, the scene is also starting.
The landing projects in the United States are mostly concentrated in the direction of passenger cars, and trial operations are still carried out on urban open roads, such as Waymo and Cruise. In the domestic market, in addition to the short-term unmanned car-hailing trial operation conducted by Pony.ai and Jingchi Technology, most players are focusing on ports, industrial parks, open roads, universities, and highways. Low-speed L4 unmanned vehicles in semi-enclosed scenarios, including low-speed unmanned delivery vehicles, unmanned sweepers, and unmanned trucks have successively landed operations, involving companies such as Zhixing, Yushi Technology, Baidu, Tucson, and Xijing Technology.
Of course, no matter what the landing scenario, the ultimate goal is to develop a car with autonomous driving technology, rather than simply providing delivery or cargo services. However, after the above-mentioned sub-scenarios are popularized, the real popularization of unmanned driving will also be accelerated.
In the competition among the crowds, transformation is a topic that has to be mentioned. Sun Guobin mentioned that traditional vehicle manufacturers have invested heavily in the field of autonomous driving, and their business models are transforming from manufacturers to car-sharing service providers. In addition, a large number of new car-making forces, namely Internet car companies, are seizing new areas such as driverless taxis and public transportation represented by Robotaxis. Both types of companies have opportunities. There is always a market for personalized cars, but shared transportation is the only way to solve the problems in big cities.
Sun Guobin finally mentioned that switching from unmanned driving to manual driving is one of the most severe challenges, which is also the difficulty in achieving level 3 and level 4 autonomous driving under all road conditions. It is a more feasible solution to limit the autopilot function to only use under certain conditions and limit the length of unmanned driving without driver intervention.
Ring Type Connecting Terminals
Ring Type Connecting Terminals,Terminals,Connecting Terminals
Taixing Longyi Terminals Co.,Ltd. , https://www.longyiterminals.com