In the 1920s, the United States, the first car equipped with an on-board radio in history slowly drove onto the road. Although the idea was not quick to popularize at the time due to the complexity of installation, high price, and suffocation of operation (driving and listening to songs cannot be carried out at the same time), the Chevrolet advertisement "Adds TONE to your car" undoubtedly announced to the world the arrival of the era of automotive entertainment functions.
In fact, the automotive industry has experienced nearly a hundred years of development by the millennium, but until the wave of smart cars came, compared with the rapid development of smart phones with more and more slippery operations, the car machine system behaved like an old stubbornness everywhere for a long time. In those years when everyone used the iPhone 4 to experience the silky fruit cutting game, the car machine could not even navigate.
Therefore, the apple Carplay with thick eyebrows and big eyes turned out to be an excellent supplement or substitute for the barren car system, and even whether a car supported Carplay in previous years can influence the purchase decisions of some consumers. Apple also said that more than 80% of new cars sold worldwide in 2020 will support CarPlay.
Is this because Apple's Carplay really smells good? The answer may be that there were too many locomotive systems that were really not so good. The short man is tall inside, and on the car machine system, Apple's piggyback hand has given the automotive industry a dimensionality reduction blow.
With the current wave of intelligent transformation of automobiles, how long can Apple CarPlay's position in the automotive industry last? The vicissitudes are too exaggerated, but the track can also be regarded as a storm.
ECUs and IVI under intelligent torrents
To understand the current situation on the automotive systems track, we need to first figure out what the automotive system really is.
I believe that many people often hear about in-vehicle OS, automotive operating system, real-time operating system, embedded operating system, and even AliOS, Hongmeng OS and other various automotive system related words. The operating system of the car does give a first impression of being diverse and foggy, and the detailed analysis of specific categories is often very complex and requires a high degree of professionalism.
Therefore, we must first make it clear that automotive systems are often not a single system, but a collection of multiple systems, and this group of systems will be more or less different because of the different classification basis.
IVI、ECU
For ease of understanding, we cut into a category from a popular perspective. Cars play two main roles in our daily use:
The first is the transportation role under the premise of safety, which is mainly regulated by the "brain" of the car. Specifically, for example, the CPU in some key parts of the car is collected, calculated, and then the operation result is converted into a control signal, thereby maintaining the normal operation of important parts such as the car engine and gearbox. These critical control units, which are essential for transport, are called ECUs (Electronic Control Units).
The second is the interaction of information and entertainment such as navigation and music during driving. This is mainly supported by the underlying system behind the dashboard, center console, car machine, etc., and the name of such a system that does not play a direct role in the driving decision of the car is IVI (In-Vehicle Entertainment Information System).
Understanding the ECU and IVI, let's further analyze the relationship between the two and the automotive system.
In fact, the ECU has similarities with the computer structure, and it is installed in important parts such as fuel vehicle engines. This means that for security reasons, the accuracy and efficiency requirements of the systems behind the support for the normal operation of the ECU are often very high.
Therefore, mobile operating systems such as Android and IOS cannot fully undertake this part of the work, and often require embedded operating systems. The disadvantages of this are also obvious, that is, the "brain" embedded in the vehicle partially limits its volume and storage capabilities, resulting in very limited hardware resources for the system, and its stored code is often only a few megabytes.
For the IVI, which is mainly responsible for information and entertainment interactions, there is no need to meet the extremely high security requirements like the ECU, which means that the "freedom" of its system is much greater than that of the ECU system, which can be achieved through mobile operating systems such as Android and IOS. Google and Apple, which focus on smartphone systems, have also benefited greatly from the ease of getting Android Aoto and Apple Carplay into the car, respectively.
The underlying system support in the era of fuel vehicles is more or less the same, ECU system solution providers mainly include Vector, KPIT, ETAS, etc., IVI system solution providers include QNX, Microltron and Google (Android), Microsoft (Win) and other technology Internet giants.
From pure driving to the pursuit of functionality, to informatization, and then to today's intelligent wave, this is also quietly changing for the system behind it.
vicissitude
In today's world where most car companies are pursuing car intelligence, new players such as Baidu, Ali, Huawei, etc. have begun to walk into this track that once seemed a bit "boring", so can the traditional ECU system and IVI system fully define these new players?
The trend of intelligence and internet of vehicles is becoming the consensus of more and more people, which means that the main battlefield of the automotive system supported by the underlying technology will gradually migrate to this in the future, and two problems that have become more prominent now are also born, one is that the iteration of the system hardware level needs to be accelerated, and the other is that the industry's binding consensus will eventually be reconstructed.
Let's first look at the current rather post-esseries, which are mainly reflected in the cause and effect at the hardware level.
Because of the surge in the requirements of intelligent technologies such as automatic driving on sensor hardware and data acquisition capabilities, in order to achieve this, it is necessary to get on the car with high-definition cameras, lidar, millimeter wave radar and other necessary hardware, which also puts forward more diverse and complex requirements for data collection, processing, decision-making and other behaviors of automotive systems.
As a direct result, the demand for the number of functional chips and the computing power of autonomous driving chips is unprecedented.
As for the industry's binding consensus, in the past, in terms of automotive safety, the industry has a very mature constraint framework in many aspects, such as the EEA electronic and electrical architecture proposed by Delphi, which not only reduces the communication cost of the industrial chain, but also reduces unnecessary trial and error risks under a consensus premise and improves the overall efficiency of the industry. This is often an objective law of an industry from emergence to maturity.
Today, we have seen many hidden dangers of safety accidents that do not exist in the era of fuel vehicles, such as accidents caused by automatic (assisted) driving by Tesla and other car companies, or many new power brands that frequently recall products due to safety hazards.
It is not difficult to find that the binding consensus of an industry will directly affect the experience of the most downstream, but the search for the reason must go back to the upstream. For example, once-'s automotive systems could be built around ECUs and IVI, it would be unrealistic to continue in the face of new technologies such as autonomous driving in the future. So how will the nascent technologies in recent years be divided into the construction of the underlying system? This is undoubtedly something to consider.
Any change and innovation must be repeatedly jumped in run-in and trial and error to reach a relatively mature position. So for the players who are already on the track, what kind of logic are they building their future pictures according to?
BAH's single point of entry competes with the full stack
Compared with the past, there have been many new faces in the automotive system track, and most of these new players are undoubtedly eyeing the most promising IVI and autonomous driving underlying system construction. Among them, those who are already in the spotlight are Baidu Apollo system, AliOS and Huawei Hongmeng OS.
First of all, Ali, as far as the current point of view, AliOS is not a supplier that can provide solutions for automatic driving systems, and in terms of intelligent transformation of automatic driving, compared with systems such as Baidu Apollo that have been put on the car, AliOS's current achievements are really not much.
As far as AliOS's circle of friends and cooperation are concerned, we all know that Ali has currently joined hands with SAIC to create a zhiji car, but even such a "own" brand has nothing to do with AliOS in terms of automatic driving, but cooperates with the autonomous driving unicorn Momenta. Tianyancha information shows that Momenta's C+ wheel is led by SAIC.
Therefore, as Xie Yan, chief architect of AliOS, said, "In the future, there are only two kinds of OS that China's auto industry will open to the outside world: AliOS, an operating system that is completely developed from the bottom of the car driving scene; and other OS based on modifications such as Android." ”
AliOS' current strategic weight in the car track is still more focused on the IVI system than on the car drive or the automatic driving system. It is IVI system solution providers such as Android, IOS, and Windows that should really be benchmarked, which is actually closely related to their strategic orientation.
Specifically, in October this year, The Zebra Zhixing, a joint venture between Ali and SAIC, released the AliOS intelligent driving system kernel, which will be provided to car companies free of charge next year, with the purpose of realizing the integration of intelligent cockpit, automatic driving and other software through the car operating system.
As an automotive system, AliOS is more like hoping to return to its own C-end business basic disk through the large-scale expansion of in-vehicle systems.
AliOS not only has cloud data advantages compared to the current mainstream mobile operating system, at the same time, Alibaba's existing ecosystems such as Taoshi, Alipay and AutoNavi Maps can jointly promote AliOS to form its own objective competitiveness in the automotive market, thus further feeding back the entire Alibaba macro ecological basic disk.
For Baidu, with the continuous deepening of the field of L4 automatic driving and Robotaxi in recent years, it is not difficult to find that the real main battlefield of Apollo is actually the underlying system construction in automatic driving.
At the same time, the artificial intelligence systems developed by Baidu such as DueerOS are likely to eventually be released in the IVI field in the form of packaging synergy with Apollo.
Compared with Ali's single point of entry, Baidu's Apollo based on the integration of many underlying systems is more like a full-stack solution to deal with the intelligent transformation of automobiles, not just for a single end point.
Starting from this logic, the basic logic of Huawei and Baidu is actually the same, that is, more strategic orientation in the early stage is to cut into the B-end market, create barriers through strong technology breakthroughs, reconstruct the entire industry chain ecology, and thus take the route of technology empowerment.
But what is more interesting is that Baidu and Huawei have distinct differences. At present, Huawei has "heavily" cooperated with several brands such as Jihu and Xilisi, not only in terms of intelligent software solutions, but also in the Lidar and other Tier1 hardware.
Baidu did not choose to participate in such a "heavy" cooperation, but personally built a car and jointly created a new brand set with Geely.
Why is it that Huawei and Baidu have similar full-stack layout strategies in automotive systems, but in the downstream practice, they have embarked on a different path?
In fact, whether it is Huawei or Baidu, the choice of the system landing path is benevolent, and the wise are wise, which is nothing more than the optimal solution modulated according to their actual situation and future strategy.
Specifically, the underlying logic of the automotive system and the automotive terminal on the mobile phone is the same to a certain extent, just like google although it has made its own smart phone terminal on a large scale, the Android system it has developed has become the choice of the vast majority of smartphone brands. And Apple has made its own mobile phone terminal, which is doomed to the IOS system to a large extent will be a relatively closed state.
Therefore, the logic of Huawei's system is like Google building Android, while Baidu is more like Apple creating IOS.
Today's automotive systems are facing the emergence of a new order after technological experience, even if the entire industry still has huge room for replenishment in terms of automatic driving, intelligent cockpits, etc., but at present, we cannot deny the fact that compared with Western companies that seized the technological opportunity dividend a hundred years ago, a number of Chinese companies are ready to go before this round of automobile change.
Reprinted from Tan Qing said AI, the views in the article are only for sharing and exchange, do not represent the position of this public account, such as copyright and other issues, please inform, we will deal with it in a timely manner.
-- END --