About the author: Procsql, a digital automobile practitioner, used to be the product manager of a foreign-funded automobile enterprise. I like reading, writing and making friends. You are welcome to discuss and exchange industry issues.
Have you ever seen the predecessor of Audi? Boss? Yes? R & ampd，？ Peter. Mertens? Articles, where? Will. Is it? Blood? we? All of them? Really? Sleep (the future will be bloody because we are really asleep), this is a very sharp article, Mertens? Experience the plight of the public in software transformation, and point out the serious and unsolvable underlying problems encountered in digital exploration.
But? Mertens? Too much emphasis on the role of software obviously does not conform to the cutting-edge cognition in the industry. The key to the future development of automobile industry lies in the overall structure. Besides software, the three elements of the architecture include hardware processor and communication module.
The article makes a metaphor: Octopus has 8 arms, and each arm has an independent brain to manage the arm. ? This? 8? The small brain communicates with the central brain. Mertens? Think of the octopus's arm as a different subsystem in the car, because? The ICE industry has matured, so each subsystem has been fully defined and outsourced to a more efficient and lower-cost one? tier？ 1? Suppliers ... real innovation has long since disappeared.
When the traditional mechanism is in a state of high perfection for a long time, the automobile industry is facing digital transformation. At this time, the waiting company has lost its creativity and can only hand over the whole product to the third-party supplier (who is also forced), while the car company does not even have the inspection ability. (Volkswagen? 90%? Will the software eventually be outsourced? ID.3？ There is a "huge software loophole"-the original official announcement of the public)
Herbert? Dees? Because? Golf? 8? And then what? ID.3？ Software disaster, lost Volkswagen Group? CEO？ However, this position has been taken over? Brandest? tter？ Doesn't turn things around. Mertens? Point it out, old man School? Professional managers have never written a line of code in their lives, or they don't understand it at all, but they are making key decisions about software.
The embarrassing situation not only troubled the public. When Mercedes-Benz and BMW face the (relatively) advanced stage of autonomous driving, both sides are dumbfounded by the new ideology, so they can only warm up first and set up R&D cooperation. Sure enough, due to unclear strategic objectives, BMW and Mercedes-Benz announced that they would suspend cooperation for less than a year. ...
Despite the difficulty of innovation, German car companies are still strong. Mercedes-Benz soon found NVIDIA to develop the next generation? 2024? In-vehicle architecture in; And BMW's? Adas? The system has been at the leading level in the industry, laying a solid foundation for the following actions. In contrast, non-German car companies are more dangerous. Many car companies didn't even struggle, either pretending to be stupid or pretending to sleep. Some car companies directly "cut the knife" for themselves and sacrificed them to start-ups of technology companies. I won't go into details here.
Peter. Mertens? The article leads to a thought: the article emphasizes the importance of software many times, so is software a roadblock to the origin of innovation? Solve the software problem, and most of the problems in the digital transformation of automobiles will be solved?
To answer this question, we first need to find out when the software will become a "thorn in the side" of car companies. A few years ago, the industry began to spread "software-defined cars". Software? Definition? Vehicle? SDV ",it is not from rigorous academic argument, but from the public at that time? CEO？ One of them? Quotation:
In the near future, cars will become a software product, and Volkswagen will also become a software-driven company. -Volkswagen Group? CEO？ Doctor? Herbert? Dis
This concept was discovered by Wang Jin, one of the "Four Kings" of Baidu Autopilot at that time, and then spread in China. Dis stands for the traditional giant, Wang Jin stands for the Internet, and the phalanx of "Big Brother Blessing", that is, the golden sentence of Big Brother, expands and spreads infinitely.
Looking back at the big environment at that time, Volkswagen put forward that "software-defined car" was limited by the times, but at that time, everyone was crossing the river by feeling the stones. Dees? Our vision and pattern have been ahead of others. But if you think about it carefully, European enterprises are too dependent on software (actually, making software focuses on operation) and have no advantage at all: they are not as flexible as local enterprises in China in demographic dividend, not as mature as Indian outsourcing in cost control, and absolutely not as good as the United States in innovation.
At that time, car companies realized the necessity of transformation, but they didn't know where to turn. Mercedes-Benz saved money from McKinsey? Case (connection? Intelligent interconnection, autonomy? Autopilot, sharing? & amp？ Service? * * * Enjoy traveling, electric? Electric drive), BMW called out these four letters in reverse order again? Ace; Instead? Dees? The strategy is pragmatic, mainly promoting new energy vehicles? MEB？ Platform, while trying to digitize the software in a low-key way.
Jack Welch (Jack? Welch) has repeatedly stressed that the strategy of the enterprise is that the boss points in one direction and the general servants run in this direction. On the other hand, four letters refer to four directions, and each direction is a bottomless pit when burning money, and it is complicated and overlapping with each other ... It can be seen that everyone was confused about the future at that time.
Li Xiang once said, "Will an enterprise like Baidu be afraid of sogou Yahoo, a search engine? No, what Baidu is really afraid of is the new media that steals traffic, such as WeChat and headlines. Similarly, are auto giants afraid of new energy vehicles? No, they will wait for the front like yellowbirds. What are people afraid of? The public is afraid that the development of the industry will no longer follow the original script. When the underlying core of the car is transformed into a highly integrated digital platform, through? Ota? When upgrading to consumers, Volkswagen does not want to follow in the footsteps of Nokia.
With the rapid iteration of the whole industry, this famous saying was quickly questioned: software defines cars, who will define the operating system? How to deal with the problem of information island in software development? How to build an ecological scene of software getting on the bus? There are many similar "killing problems", which are practical obstacles that product managers encounter in their work. If you look for the answer from the software level, there is no solution at all.
The main task of the car is to travel, and passengers have to get off at their destination. Whether there are unfinished songs in the car or short films in the middle, it is a side task. This makes it difficult for popular applications such as "eating chicken in the car" to be displayed on the mobile phone.
Starting from the car scene, we can make good products. Take the relatively mature car map software as an example, most products still show obvious castration of their mobile phone functions. Is it because the product manager is not focused enough? Of course not. Digging deep into the reasons, there is no convincing closed-loop business at present, which makes developers willing to invest real money in car-side application software.
The fragmentation of the scene inside the car is not "super?" ID "or" family bucket "and other software application levels can be solved. Issues related to the nature of business are all strategic issues. Problems solved at the software level, such as user interaction, such as UI design, are superficial tactical problems. History tells us that we can't use tactical diligence to cover up strategic laziness.
The higher the position of the digital module in the vehicle architecture, the more the on-board software will get out of the information island. "Software" is not enough to define the characteristics of converged applications. Software iteration is getting faster and faster, and the demand for hardware configuration behind it is also increasing. In other words, software often solves superficial problems. If hardware and communication can't keep up, we must first solve deeper and more basic problems.
The basic problems are roughly divided into two categories: the lack of computing power of hardware chips in a narrow sense and the lack of communication capabilities in a broad sense. Where there is a problem, someone will try to solve it. Companies such as Qualcomm, NVIDIA and Intel are marching into high-computing chips, and "infrastructure fanatics" such as Huawei and the three major operators are solving the problem of communication facilities.
The key for non-car-code chip giants to enter the closed and extremely high-barrier automobile chip market is to find a balance between product performance, quality and cost. Quality takes time to test, and chip companies have really made great efforts in product performance and cost:
Han Xiao, editor-in-chief of Auto magazine, wrote an article introducing the performance advantages of Qualcomm chips? ——? "The new model can be said to have several tablets installed in the car; The demand for in-car entertainment is increasing rapidly, and the demand for computing is also increasing ... In the face of complex task processing, the performance of classic car chips is tight; Compare Qualcomm? 820A？ The performance gap between the chip and the traditional automobile chip can be seen very intuitively (regardless of the process or? CPU？ Main frequency, etc. ), not to mention? 820A？ It is the only chip on the market that supports "one core and four screens".
Compare the product technical specifications of Qualcomm and several traditional automobile chip manufacturers (source: automobile things watch)
In terms of cost, "car-level chips" can't be counted. Cars often only need one or two high-power chips. What is the number of cars? What is the number of mobile phones? Who doesn't first calculate whether the input cost is amortized on a large scale? In fact, more and more mainstream car companies are beginning to abandon car regulations and directly use consumer-grade chips. A few months' development cycle is enough, the performance has been severely tested by mobile phone users, and it is durable, and the cost is amortized by more than half. It also avoids the car rules and? ASIL？ Authentication requires a lot of money and time. The car regulation threshold itself is the barrier of traditional car companies, which is not discussed in this article.
Some people worry that consumer chips can't meet the working life requirements of vehicle regulations for temperature environment (-40℃)? 125℃)。 However, extreme temperature environment is rare in reality ... why should the cockpit be required to reach the extreme temperature when passengers can't stand the temperature?
In-vehicle communication is a further challenge to automobile digitalization, and its return is also rich: 5G? Technology will greatly change the intelligent cockpit, driving assistance and so on. And bring a full range of experience upgrades to the travel ecology.
The software of the intelligent cockpit needs to be upgraded. Now it is not a "stand-alone version" that can meet the needs of consumers. A car with an upgradeable system is a "life". However, the current vehicle transmission capacity barely supports the foundation? OTA, it is difficult to meet higher demand, such as? C-V2X？ The described interconnection between vehicles and other terminals requires a "C-V2X" communication system with high bandwidth, low delay and high reliability.
Are you online? 5G？ Before the communication technology is mature, the driving assistance scheme is realized by the sensor's perception of the surrounding environment. Sensors on the vehicle, whether millimeter-wave radar or camera, need a series of preconditions to ensure the safe driving of the vehicle: closed-loop road conditions, low-speed driving, and extremely high computing power redundancy of the processor.
Some insiders compare the sensor to a guide rod, which is by no means the eye of a car; Imagine a person walking with his eyes closed and using a guide stick. In order to avoid colliding with obstacles, he can only walk slowly and dare not run, and his brain should also leave extra space to analyze complex road conditions.
I'm more nervous than it drives? This is the real feedback from countless drivers who have experienced the driver assistance system.
This experience is transitional. What will happen in the future? 5G？ Commercialization has become history: high-precision maps can play a greater role, such as providing God's perspective in real time, and even allowing vehicles to communicate with other intelligent facilities. It will completely change the technical scheme of driving assistance, and truly achieve "communication positioning is the mainstay, supplemented by perception scheme"; In the meantime, for by going up one flight of stairs? L4？ Autopilot provides a technical path.
Putting the future of intelligence on software is a bit like the westernization school's "learning from foreigners to control foreigners" a hundred years ago. The backwardness of the late Qing Dynasty is not a technical problem, but the cradle of agricultural civilization's failure to provide science and technology. The same is true of the transformation and digitalization of automobile enterprises. On the surface, it is a software problem, and the core is the backward infrastructure.
Elon? Musk? Development? Models? s？ Control the wiring harness of the whole car when obsessive-compulsive disorder occurs? 3? Within a few kilometers. What happened afterwards? Models? 3? What are the harness requirements? 1.5? Kilometers, model? y？ what's up 100? rice. This move was forced? Tesla? Step by step, right? EE？ Cut and optimize the network architecture and cable topology, eliminate those that can be eliminated, integrate those that can be integrated, and use those that can be transmitted wirelessly.
At that time, I was the driver assistant product manager of a foreign-funded automobile company. I asked R&D by email, "How long is our next-generation fist product?" The answer is short and powerful: "Why? Who is it? Care? Talking about Tesla's products, some people see the rough appearance and interior, while others see the ultimate works of art in mechanical aesthetics, and there are different opinions.
This article comes from car home, the author of the car manufacturer, and does not represent car home's position.