GAC's on-going development of intelligent connected cars

* Trumpchi GS4

For a long time, my country's automobile industry has been in the "production and manufacturing" stage, and joint venture brands have occupied the main market. However, in recent years, more and more domestic brands have begun to rise, and my country's independent automobile research and development capabilities have gradually increased.

Among them, GAC Group established the GAC Research Institute in 2006, which attracted many outstanding talents from home and abroad to plan and develop new products and technologies. Through independent forward development, the GAC Research Institute has a relatively complete vehicle and machine development capability. As of January 2018, 18 vehicle products in two series have been developed and launched, and the two popular SUV models of GAC Trumpchi GS4 and GS8 have been successfully created.

* Trumpchi GS8

With the rise of the intelligent connected car industry, GAC has gradually paid more attention to this field and started independent research and development. Faced with the emergence of new car-making forces in China, they are also actively embracing Internet thinking to improve the user experience of their products.

In 2013, GAC Research Institute released its autonomous driving new energy concept car Witstar, which has now developed to the second generation. In June this year, the "Qiyun" concept intelligent connection system jointly developed by GAC Trumpchi and Tencent was also officially released.

*GAC's second generation driverless car, Qiji WitStar Ⅱ

"Relatively speaking, the gap between China's IT industry and the United States is not that big. From the perspective of the integration of intelligent connected technology, through the integration of different domestic industries, perhaps in this field, China's independent research and development speed will be faster." said Shang Jin, general manager of GAC Silicon Valley R&D Center.

In April last year, GAC registered and established a research and development center in Silicon Valley. "Through this platform, we can access the world's forward-looking technologies, recruit global automotive talents, and exchange information on a global platform," said Wang Qiujing, director of GAC Research Institute, in an interview with the media.

According to Leifeng.com, nine automakers, including Ford, Volkswagen, GM, Mercedes-Benz, BMW, Toyota, Nissan, Honda and Hyundai, have already established R&D centers in Silicon Valley. As a gathering place for the world's top high-tech talents, Silicon Valley has become a place where companies related to intelligent connected vehicles are vying to take root.

"We now have 18 PhDs, and I am the only 'local PhD', while the rest are graduates from the top 10 universities in the US. We have worked for an average of nearly 10 years," Shang Jin said with a smile. But these words cannot hide his unique background: he obtained his bachelor's, master's and doctoral degrees from the Department of Automotive Engineering at Tsinghua University. His doctoral dissertation on tire mechanics won the "National Excellent Doctoral Dissertation" in 2003, which is the only one in the automotive industry so far; he is currently also a visiting researcher at the Department of Automotive Engineering at Tsinghua University. After graduating with his doctorate, Shang Jin went to the United States and switched to the IT industry. He has been engaged in network and security in Silicon Valley for 16 years and has more than 10 US patents in the field of network security.

*Shangjin

At the 2018 International Forum on Autonomous Driving Vehicle Safety Technology held by SAE this year, Leifeng.com New Intelligent Driving conducted an exclusive interview with Shang Jin. The following is the transcript of the interview, which Leifeng.com has edited without changing the original meaning.

New Intelligent Driving: When did you start to join the automotive industry?

Shang Jin: I joined GAC Silicon Valley R&D Center in June last year. I was studying automobiles. Nowadays, the automotive technology industry in Silicon Valley is not mainly focused on developing traditional automotive technologies, but solving automotive problems from an IT perspective. This can be said to be an innovation and change. This is also the reason why a large number of IT talents, or people from different backgrounds, are engaged in this industry.

New Intelligent Driving: What areas does GAC Silicon Valley R&D Center mainly focus on and what is its positioning?

Shang Jin: We set up the Silicon Valley R&D center in the hope of attracting more high-end talents and conducting external cooperation. Our goal is to build platform technologies for vehicle manufacturers, which require the integration of IT and traditional automotive technologies. Specifically, we also refer to the development routes of many automobile associations, industry academic institutions, and Silicon Valley high-tech companies, and set several dimensions: autonomous driving, big data analysis, information security, and new energy powertrains. The first three are related to intelligent networking, and the last one is related to electrification.

Autonomous driving is a platform technology. Only with such a platform can the OEM play a leading role. In the future, autonomous vehicles will become a new data source, just like search engines more than ten years ago and mobile phones five or six years ago. These data will promote the creation of a new economy or change the existing economy. This will increase the value of the car, but it may also attract hacker attacks. Therefore, autonomous driving, big data analysis, and information security are also our main areas of focus.

We mainly focus on self-research, but we will also make good use of local resources in Silicon Valley and integrate the technology of GAC Research Institute in Guangzhou. Because the IT industry is different from the automotive industry, seven or eight engineers can set up a technology company, but in the automotive industry, the development of each major function will not work without a team of at least 50 or 60 people. It will cost a lot to set up a team of hundreds of engineers in Silicon Valley. Many things can be done well in China. The Silicon Valley R&D Center plays the role of a leader, leading more people to realize the implementation of technology together.

New Intelligent Driving: What are the professional backgrounds of the team members?

Shang Jin: All members of the team are engineers recruited from North America or Silicon Valley. Their professional backgrounds include autonomous driving, vehicle majors, and Internet algorithm backgrounds. But overall, there are more talents in software. Currently, there are more than 10 people in our team who graduated from Tsinghua University, two of whom studied in the junior class, and one of whom was the top scorer in the college entrance examination. Our average working time is nearly ten years, and most of them come from Silicon Valley high-tech companies, such as Cisco, AOL, Yahoo, etc.

In addition, we also collaborate with American universities CMU (Carnegie Mellon University) and UC Berkeley (University of California, Berkeley) to research more forward-looking technologies.

New Intelligent Driving: What is the current status of your autonomous driving research and development?

Shang Jin: We are now promoting the mass production of L3 models, and hope to launch them on the market within one or two years. GAC has already set a clear timetable. In addition, we are also considering and promoting L4 autonomous driving. Our mass production time is not slower than similar products on the market.

New Intelligent Driving: In terms of hardware and software, do you have some products developed by yourself, or are you integrating them from Tier 1?

Shang Jin: We will definitely cooperate with Tier 1 and more suppliers. Vehicle manufacturers are OEMs, and they will gather platform technologies and will inevitably use a lot of Tier 1 technologies. At the same time, some key technologies are also what we, the vehicle manufacturers, must lead and master. In addition, if we do a good job on our own platform, good Tier 1s will also be attracted. This is our idea.

New Intelligent Driving: What do you think are the difficulties in developing autonomous driving?

Shang Jin: It is different at different stages. From the perspective of L3 autonomous driving, I personally think that from the perspective of some current products, the difficulty may not lie in local demand, but in the system design and integration capabilities. Of course, there are also safety and cost considerations behind it, but the safety cost actually reflects the system equipment and system integration capabilities. This is the difficulty for vehicle manufacturers to promote L3 autonomous driving now or in the next one or two years.

It is still hard to say what L4 autonomous driving will be like. The current difficulties mainly come from some technical bottlenecks, such as how to integrate perception and planning with AI in a more complex environment. In addition, some basic things are not yet mature and need to be well planned. V2X may be able to make a breakthrough quickly and provide help. But this is not a technical difficulty of autonomous driving technology, but a hope to form an ecosystem.

*Shangjin gave a speech at the SAE 2018 International Forum on Autonomous Driving Vehicle Safety Technology

New Intelligent Driving: How to ensure the safety and network security of autonomous driving cars?

Shang Jin: Automobile safety actually refers to functional safety, which is a very complicated issue. If this issue is resolved, autonomous driving has actually been completed. To achieve the best automobile safety and high reliability, the integration of the two major industries of IT and automobiles is still needed.

The integration here includes the integration of technology and methodology. But which one should be the main one in the integration? It is hard to say today, but I believe that we will reach a conclusion after working together. In fact, the integration of methodology is the integration of people's concepts.

With the deepening of intelligent networking and the increase of processors and software inside vehicles, information security will also be very complicated. The integration here will be a little more difficult. We can learn from the methods of implementing information security in the IT industry, borrow its practices and make some modifications. Because the network information security issues and protection methods in these two industries may be mostly the same. Of course, it can also be redesigned. Maybe in general, the two industries are the same, but they are different when it comes to the car, and need to be redesigned.

Therefore, there are two main difficulties in network information security. One is to distinguish which things can be directly used in the automotive ecosystem, such as cloud security issues of vehicles, which have nothing to do with the car itself. Another difficulty is that many things need to be processed or redesigned, and we can only borrow ideas.

New Intelligent Driving: In the future, cars and sensors will generate a lot of data. What value do you think this data can generate?

Shang Jin: There are two aspects. On the one hand, both big data and AI can be used in all industries, using massive amounts of information and data to solve problems in almost all industries that are currently difficult to solve with traditional physical models.

Using big data analysis in cars can improve the quality of cars and enhance the user experience.

Specifically, new energy batteries have been tested a lot before vehicles leave the factory, but today, they are far from covering real scenarios, and batteries may be very sensitive to these, so they can be improved through big data analysis. After the owner buys the car, the battery data is accumulated through monitoring and big data analysis, just like a black box system, and finally the battery life is improved through diagnosis, prediction and cloud control, which improves the user experience. This is also a way to help improve the quality of the car through big data analysis.

From another perspective, using in-car data for big data analysis, especially in new market areas such as UBI (user-based vehicle insurance), is a new "cake". In addition, big data analysis will redivide the existing "cake", such as the 4S store system. Both of these "cakes" are worth paying attention to in the future.

New Intelligent Driving: What do you think of some of the new car companies nowadays?

Shang Jin: We welcome competition, which is good for everyone. Without new car companies, traditional car companies will not seize the opportunity to change. The emergence of new car companies can also help us discover our own problems and give full play to our own advantages.

Google doesn't do evil, Amazon and Microsoft employees: neither should we

Leaving Google and returning to Stanford, the new move of "AI Goddess" Fei-Fei Li is revealed

Things Huawei didn’t tell you at the Mate 20 launch event