The cars that have changed the world are being changed, and the "new four modernizations" of automobiles are giving birth to major changes in the automotive industry in the past century.
From being defined by hardware to being defined by software, the attributes of cars are also quietly changing, from traditional means of transportation to the "third space" evolution, consumers' focus is also shifting to ride experience, intelligent network configuration. There is no doubt that in the future, the car will evolve into the largest intelligent terminal, a third space integrating entertainment, travel and office.
On September 27, in the "Titanium Media Parlor", Liu Xiangming, co-founder of Titanium Media Group and founding director of ITValue, and Dr. Zhang Lin, founder and chairman of Forteca, launched a dialogue on "software-defined cars", "This is not only a fashionable topic in the circle, but also an enlightening topic, and indeed gradually affecting the fundamental changes in the automotive industry." Zhang Lin said.
The "Titanium Media Parlor" column was created by Titanium Media during the World Internet Conference, which is the only "science and technology media live broadcast room" built by the "World Internet Conference and Internet Development Forum" - Titanium Media Live Broadcast Room.
<h2>Software-defined automotive, SOA-based service architecture</h2>
At present, in the context of software-defined vehicles, service-oriented software architecture SOA is rapidly being borrowed from the IT industry to the automotive industry, and some of the processes and practices that originally defined hardware have been rapidly transformed into some new technical concepts such as defining software and defining services, from implementing product definition, development and verification.
However, when cars can be software-defined, the hardware hidden behind them will be highly convergent. In this regard, Zhang Lin said that the differentiation of automobiles will be mainly reflected in the innovation of software and algorithms, which will reshape the entire automotive supply chain system. He believes that as the ecology of the car, the software development platform with SOA as the main body will become the industrial growth point of intelligent cars.
Intelligent connected vehicles are transforming the entire automotive industry chain from the past vertical integration model to a round table, or a chain-like cooperative ecological model, and each player will find new value in this new value chain.
Regarding the core value and implementation logic of software-defined cars, Zhang Lin said that first of all, the software implementation function will be more cost-effective than hardware, which is one of the most basic values of software-defined cars.
Secondly, software-defined cars will bring a lot of value-added value, and will also change the marketing model of the entire automotive industry, and car sales will no longer be the end point, but become a new starting point for serving customers. When a vehicle is delivered to the consumer, through the OTA (vehicle software upgrade) upgrade, after-sales can continue to obtain a lot of value-added services. Car companies and suppliers along the value chain will accompany the whole life cycle of the product 7× 24 hours to serve the terminal customer.
The third is the synergy value, the software has its flexibility, scalability, can be transplanted and reused in different products, some high-value such as hundreds of thousands of vehicles on the function, we can also be reused on tens of thousands of vehicles, this is the charm of the software. In addition, through the software architecture, it is more flexible and efficient, and can realize the collaborative development of the software ecosystem of the main engine factory and multiple partners.
Therefore, through the practice of software-defined cars, Zhang Lin believes that the future is more likely to move towards a thousand people than to the past.
Today, there could be 100 million lines of code on a luxury car, far more than the number of codes on Boeing planes, Airbus planes. In Zhang Lin's view, this trend is still accelerating, and the number of codes on the economy car in the future may have hundreds of millions of lines of code. When the software is embedded in the vehicle and used, the software becomes an integral part of the vehicle, and once the OTA is equivalent to redefining the vehicle, the car will become a completely new product, which is why the smart car has the ability to grow.
Liu Xiangming, co-founder of Titanium Media, spoke with Dr. Zhang Lin, founder and chairman of Fortek
Recently, the Ministry of Industry and Information Technology has issued relevant new policies for smart car manufacturers and products, in which the software upgrade (OTA) of the vehicle is frequently mentioned, for example, the car company needs to establish a management system and guarantee mechanism for the software upgrade of the vehicle, and when the car company implements the OTA upgrade, the technical parameter changes involving safety, energy saving, and anti-theft must be reported.
Zhang Lin believes that because cars have extremely high safety, durability, and reliability attributes, and any line of code can theoretically change the attributes of the vehicle, it is necessary to be rigorous and cautious about OTAs. For an OTA product, once it is recognized as a new product, it must undergo a rigorous and careful verification, testing, and release process. The OTA requirements of the car are not the same as the OTA requirements of the mobile phone, after all, the vehicle safety attribute is the first priority.
As the proportion of software becomes larger and larger, architecture will become more and more important, and car companies must fully consider all aspects of design, research and development, and manufacturing. Zhang Lin believes that in the future, in the new architecture of the car, there are functions of cloud, vehicle central computing platform, intelligent gateway, and data communication, and will also undertake some computing functions, which can support the implementation of the three major functions of the intelligent control platform, the intelligent driving platform, and the intelligent connection platform.
Foretec has a flexible and open cooperation model, such as in some vehicles, through VDC or intelligent control platform, to achieve some ADAS functions. By providing complete and mature algorithms, the software modules are integrated into the oem's existing VDC platform and thus used in conjunction with some traditional functions. This process involves working with the relevant partners, as well as the OEMs, to provide modules with their respective functions, leading to the complete delivery of the vehicle throughout the VDC.
In terms of moving toward autonomous driving independent domain controllers, the modules of the software architecture will be very complex, including services, requirements definitions for scenarios, APIs, middleware, OS, underlying software, and virtualized hardware abstraction layers. In addition, in high-level automatic driving, cameras, millimeter wave radar, ultrasonic radar, and lidar will be more and more used, and driving and parking will be highly integrated. In this huge system, the automotive architecture will become more and more complex, and the ability of various partners to work together will be more and more tested.
<h2>
Intelligent driving system, the consideration of safety attributes is the first priority
</h2>
At present, the development of automatic driving technology is in an embarrassing period, the artificial intelligence behind today's automatic driving assistance system can not reach the intelligence that people have, so in some scenarios, we see that the automatic driving assistance system is a mechanical judgment, and many accidents often occur on the interface of scene switching.
Intelligent driving, or now the so-called high-end driver assistance system, is still based on the driver, and the system helps the driver to drive more safely and comfortably. When the driver is misjudged or not careful, the intelligent driving system can help the vehicle to brake safely and urgently. In many scenarios, the safety of intelligent driving systems has been reflected.
At the Hainan New Energy Vehicle Conference, Tesla CEO Musk said through a video speech that in the future, autonomous driving can be achieved through visual neural networks, which is more than 10 times safer than ordinary people driving.
Zhang Lin also believes that in the future, automatic driving will be safer than human driving, because most traffic accidents are caused by human factors, such as people getting tired while driving, napping, deserting, etc., which will bring hidden dangers of safety accidents. But the computer system will not, its biggest feature is accurate execution, from this point of view, a good driving system will indeed be safer than human drivers in the future.
However, it is worth noting that when the autonomous driving assistance system can only make a good judgment through the scene of deep learning, once it jumps out of the scope of the scene it has learned, it does not have this inductive reasoning logic of the person, and it may make a greater chance of making mistakes.
Zhang Lin believes that under different scenarios, the automatic driving assistance system should make different judgments, how to correctly use intelligent driving or intelligent driving in those scenarios to be safer, these are the need to let the public have a correct cognition.
In his view, for the automatic driving assistance system, the consideration of safety attributes is the first priority, followed by how to bring a more relaxed and comfortable driving experience.
In terms of safety design, The Fredtec solution includes the basic elements of functional safety ISO26262, SOTIF expected functional safety, and information security. At the same time, as Tier1, Fredtek provides a high-performance, scalable hardware platform while undertaking the implementation of software modules. In this process, Fredtek collaborates with OEMs and ecosystem partners to define, develop, integrate, validate and test the software development process.
Since its inception in 2016, Fredtek has continuously promoted the evolution and mass production of ADAS and autonomous driving localization technologies. In the past few years, Foretek has cooperated with more than 40 OEMs, and its software and hardware solutions have been fixed in the front-loading of more than 40 models covering commercial vehicles and passenger cars, including many explosive models such as Changan UNI-T, and 400,000 sets of domestic vehicles are expected to be equipped with its products and systems this year.
The technical route of autonomous driving, choosing a progressive development model
In the field of intelligent driving, in addition to startups like Fredtek, there are also various technology companies that are also cutting across borders, and even car companies have begun to develop themselves, and the competition is becoming more and more fierce.
Zhang Lin said frankly, "When you are faced with the different tastes of multiple customers, how to efficiently equip different dishes in your kitchen to meet the needs of each family may be an advantage to form in the competition." ”
Frettec is rooted in the Chinese domestic market and faces the needs of OEMs: on the one hand, around the general trend of soft and hard decoupling, it provides more hardware scalable and software-upgradable customized services to OEMs with platform advantages.
On the other hand, with the support of mass production experience, Fredtek has more than three years of front-loading mass production experience in the integration of camera and radar perception and the understanding of China's road scenes, so as to ensure the development and industrialization of its products.
According to Zhang Lin, in mass production, Fortex solves a large number of complex, long-tail problems for China's road conditions, which can ensure the safety of automatic driving; mass production brings a large number of vehicles running on the road, which can really enable Fredtek to accumulate data and the algorithm is constantly upgraded; mass production is also a process of engineering iteration, Fredtec not only provides advanced systems and software, but also focuses on engineering iteration with a lot of energy.
In the choice of autonomous driving technology route, Fredtek believes more in the progressive development model. Zhang Lin believes that the gradual development model is a more desirable direction both commercially and technically, when the product can achieve mass production, there will be revenue and profit, and with hematopoietic ability to support the further development of enterprises.
The field of automatic driving is a very long-term investment, Zhang Lin believes that gradual development is easier to show the ability of enterprises to return on investment, complete unmanned driving sooner or later will come one day, but how long this cycle is, no one can accurately predict.
In Zhang Lin's view, the investment heat of capital in the field of automatic driving has not only not cooled down recently but has increased investment. In the past two years, capital is very optimistic about the overall track of autonomous driving, but in the specific investment direction, the degree of attention has begun to change, and it has begun to pay attention to the return on investment and whether the product can be mass-produced, which is why many autonomous driving companies have begun to emphasize their mass production landing capabilities.
The road to intelligent driving is very long, and continuous development requires continuous investment, which all depends on funds. According to Zhang Lin, in the first half of this year, Fredtek has just completed more than $100 million in Series A financing, and is also planning to open a Series B financing in the fourth quarter of this year.
For many companies at home and abroad this year have begun to focus on the autonomous driving freight track track, Zhang Lin believes that many new players cut in, there are different reasons behind this, on the one hand, RoboTaxi due to the higher requirements for safety of manned; on the other hand, driven by capital, hope to see commercial application scenarios, and the freight track this scene is relatively speaking, may be realized faster.
Whether it is intelligent driving on freight trucks or intelligent driving on passenger cars, Zhang Lin believes that there is not much essential difference in the entire intelligent driving system and architecture, but there will be differences in the specific perception, planning, decision-making, control and other levels of details, such as in terms of brakes, commercial vehicles and passenger cars have different requirements for perception; in addition, the commercial vehicle body is relatively long, the blind area is relatively large, and the deployment of sensors is different from that of passenger cars.
At the same time, it is worth noting that there are also different vertical scenes in the freight track, such as heavy truck trunk logistics, ports, mines, etc., and even express delivery vehicles at the end.
All of these autonomous driving subdivision vertical scenarios have a common feature, that is, landing faster. In these aspects, Fortex is already the leader of the domestic front-loading commercial vehicle ADAS, and has become a supplier of many commercial vehicle companies, including Dongfeng, Foton, Foton Daimler, Shaanxi Automobile, Dayun, etc. Forteca are providing mass-produced ADAS and automatic driving system products.
In the process, Fredtek is also upgrading its products from L1, L2, to L2+ to L3. In Zhang Lin's view, the application scenarios of the autonomous driving freight track are very wide, the potential value is very large, the Chinese logistics market is a huge cake, I believe that every participant has the opportunity to share a piece of cake, even if it is a small piece is very attractive.
In short, in the era of software-defined cars, the industrial chain of smart cars is very long, the boundaries of automatic driving can be extended and expanded are very wide, and every player in it, as long as there is excellent product and technical strength, will be able to ride the wind and waves in the wave of "new four modernizations" of automobiles.
(This article was first published on titanium media APP, author | Zhang Min)