As an extremely critical part of the AI industry, intelligent travel has become an increasingly concerned area from the industry to the public. In this context, the well-known AI information platform Machine Heart created Auto Byte, a platform focused on the smart mobility industry, last year. On March 23, Auto Byte held the "Chief Chi Xing Conference" based on its high-end interview brand "Chief Chi Hang Officer", inviting 11 well-known industry leaders to express their weighty and representative views on the current situation and future of their respective fields.
This "Chief Intellectual Officer Conference" gathered a number of leaders from vehicles, vehicle-grade chips, and multi-category autonomous driving companies, who are not only top technical experts, but also have unique observations and experiences on the development of the industry.
Xia Yiping, CEO of Jidu Automobile: Automotive robots open the era of intelligent car 3.0
In the view of Xia Yiping, CEO of Jidu Automobile, when the car has the ability to move freely, communicate naturally and grow itself, it is actually a robot with the form of a car. The car robots that focus on intelligent travel and are injected with AI will open the era of intelligent car 3.0.
According to Xia Yiping, the computing power of the previous vehicle-grade chips has long been lower than that of consumer-grade chips, resulting in AI technology not being able to play an advantage in the car, but the smart car 3.0 era can give the car enough computing power to gradually transform it from a means of transport to an AI-driven intelligent mobile space, which in turn brings technological innovation, efficiency improvement and experience subversion.
"We believe that 2023 will be the first year of intelligent competition for automobiles, and the real era of automobile 3.0 has arrived." Xia Yiping judged this.
Xia Yiping said that the intelligent car 3.0 era will also pay more attention to the safety and stability of software, in addition to relying on the protection of high-level hardware, it is also necessary to work the software algorithm capabilities, through the accumulation of safe mileage data, and continue to improve the capabilities of the automatic driving system.
Unlike other brands, Jidu Auto can apply Baidu Apollo's high-end autonomous driving capabilities and safety system in the full stack, which is also one of its main advantages. The latter currently has 25 million kilometers of safe and zero-accident autonomous driving road test mileage, and has continuously carried out real road tests in more than 30 cities across the country.
However, due to the different practical application scenarios of the two, Jidu Automobile also needs to be re-integrated and developed on the basis of Apollo. To this end, Jidu Automobile has developed a set of research and development processes exclusive to the brain of automotive robots - SIMUCar software integrated simulation prototype, which can make the software and hardware development of the whole vehicle not interfere with each other, and put the development of automatic driving capabilities in advance, so as to verify the safety and stability of the software and obtain more sufficient time.
Gu Weihao, co-founder and CEO of Zhixing: Data intelligence is the most fundamental driving force for the evolution of autonomous driving AI
In the field of autonomous driving, test mileage and test scenarios are important factors in determining the capability and safety of automatic driving systems.
Gu Weihao, co-founder and CEO of Zhixing, said that data intelligence is the most fundamental driving force for the evolution of autonomous driving AI, and the better algorithms and service models obtained by further learning and mining training on feedback data can bring better system performance to users. In this process, cost and speed are the two most critical aspects, and they are also the ideological seal of data intelligence.
Based on this thinking, last year, Zhixing launched China's first data intelligence system dedicated to autonomous driving, the MANA Snow Lake System. The system includes four subsystems: TARS (Data Prototype System), LUCAS (Data Generalization System), VENUS (Data Visualization Platform), and BASE (Underlying System), which can realize data acquisition, expression, storage, transmission, calculation, and verification, thereby reducing product costs and improving iteration speed. In the future, the system will also serve as the underlying system of the terminal AI and become the core driving force for the evolution of data intelligence.
According to Gu Weihao, "All of these capabilities will be concentrated on a recent product , Small Box 3.0, which is expected to be mass-produced in the first half of this year." Next month, Chi Heng will also host the next AI Day, when further SOP practice plans and then-competence performance will be announced." Up to now, the mileage of assisted driving users has exceeded 6 million kilometers.
In terms of longer-term planning, Zhixing has also formulated a product roadmap for the next year: in the first half of this year, the city NOH, that is, HPilot3.0, will be launched, and the system will be further iterated in the second half of the year to achieve full-scenario access. Next year, HPilot 4.0 with more assisted driving capabilities will be launched, and there will be a new product called HSD (HAOMO Self-Driving).
Wang Ping, CEO of Cambrian Xingge: Bicycle intelligent breakthrough, cloud-edge end car collaboration
As a chip company focusing on the field of autonomous driving, Cambrian Xingge has deep insights into the autonomous driving and chip industry. Wang Ping, CEO of the company, talked about the multiple challenges faced by the large-scale landing of intelligent driving on the chip: the single-chip computing power is not enough, so it needs two or even multiple pieces to achieve, but this leads to a significant increase in system complexity and power consumption, increasing the cost of the system, making it difficult to popularize it in fuel vehicles or economic electric vehicles below 100,000 yuan.
When it comes to the future trend of autonomous driving chips, Wang Ping also gave two judgments: "One is universal openness, and the other is large computing power."
Wang Ping said that in the era of L1 and L2 level automatic driving, because the amount of data is relatively small, many car companies can accept the closed integration scheme of strong coupling of chips and algorithms, but the amount of data in the L3 and L4 eras has surged, and the algorithm is more complex, and it needs a large computing chip to meet the demand.
In the future, Cambrian Will also launch products covering different levels of autonomous driving, including SD5223 (this year) and SD5226 (next year) that will be launched this year and next year. Among them, SD5223 is a product for the L2+ market, the maximum computing power exceeds 16 TOPS, and a single SOC can realize the function of the integration of row and berth; SD5226 is for the L4 market, supporting the end of the car training products, using 7nm process, AI computing power of more than 400 TOPS, CPU maximum computing power of more than 300K + DMIPs.
In addition, vehicle-road cloud collaboration will also be an indispensable part of achieving high-level automatic driving, but there are also challenges in this regard: the closed loop of massive data needs the support of large-scale AI clusters, not only the cost pressure is relatively large, but car companies also need to invest huge resources to achieve data security and privacy protection, as well as to meet the personalized needs of car owners.
According to Wang Ping, Xingge Technology will cooperate with parent company Cambrian to launch an integrated plan for cloud-edged end cars. In the cloud, with the help of Cambrian's existing high-performance training chips, the trained model can be pushed to the car end through OTA; at the edge end, the intelligent chip based on the Cambrian edge end and partners can launch a road test unit for vehicle-road collaboration; at the car end, through the SD5226 and other automatic driving chips, support the demand for large computing power of high-level autonomous driving complex models in the future, as well as the continuous iteration of algorithm models.
Li Bo, Vice President of Lotus Technology and Head of Intelligent Driving Business Line: Continuing lotus genes with "track-level autonomous driving"
When the era of electrification and intelligence brings opportunities to the automotive industry, it also means that traditional car companies can no longer follow the rules. As a racing and sports car brand with a glorious history of 74 years, Lotus has officially opened an innovative exploration of a new track after the clear transformation to electrification.
Last year, on the occasion of the establishment of Lotus Technology, Mr. Feng Qingfeng, CEO of Lotus Group, set a ten-year goal of "track-level intelligent driving" technology to beat the F1 championship. At this conference, Li Bo, vice president of Lotus Technology and head of the intelligent driving business line, explained for the first time the deeper reasons behind this goal.
In Li Bo's view, the ability of intelligent driving systems has begun to replace power performance and become the most critical part of the car. Lotus hopes to achieve the level of meeting the requirements of competition by creating a "track-level intelligent driving" - with a higher precision full coverage of the perception ability, a better understanding of the game, faster and more stable control ability, to create Lotus's unique "track-level" intelligent driving.
In addition, Li Bo also stressed that Lotus's goal is to create an end-to-end intelligent driving experience that can cover high-speed expressways, urban roads and parking scenarios. Li Bo also took the mobile phone as an example: the application and experience of iPhone 1 and iPhone 13 have not changed in essence, the difference is only in performance, storage and size, and the intelligent driving system can currently achieve the same end-to-end full-scene coverage as the future.
In this era of intelligence, compared with the traditional automatic driving classification system and the change of function to performance, Li Bo believes that in this process, "takeover mileage" and "coverage mileage" will become more important core indicators for future smart cars.
At the same time, Lotus will also build its own intelligent hardware, intelligent software and intelligent cloud, Li Bo said: "Software-defined cars, hardware-defined software ceiling, and intelligent cloud is the productivity factory of intelligent driving, which will greatly improve the speed of software update iteration of intelligent driving." With the same hardware ceiling, with the smart cloud we'll be able to reach that ceiling faster."
He also proposed that in the current evolution of intelligent driving, the current historical stage needs to pursue more, higher, faster and stronger technologies. On the first pure electric SUV TYPE 132 or more production models of Lotus to be unveiled on March 29, create a stronger perception system, a stronger computing power platform, and a more responsive and redundant electronic architecture.
Jianxiong Xiao, Founder and CEO of AutoX: How does RoboTaxi really commercialize?
For another route for the commercialization of autonomous driving, that is, Robotaxi, which is regarded as the final application scenario of autonomous driving, the industry has also maintained a high degree of attention for a long time.
Xiao Jianxiong, founder and CEO of AutoX, said that although autonomous driving can have a variety of applications such as sweepers, low-speed minibuses or unmanned ports, only by using this technology as a major change can it be possible to produce real breakthroughs and bring huge changes to human life.
AutoX has always focused on removing the safety officer's L4-level unmanned RoboTaxi, and Xiao Jianxiong also believes that this is the only way to truly commercialize the route: only by achieving the same practicality as existing online ride-hailing vehicles, and completely removing the safety officer, unlimited destination, and unlimited area of autonomous driving, is the real commercialization.
Among them, the area covered is the most important point for Xiao Jianxiong. He said that RoboTaxi's commercial operation must have a large enough service area, if it can only run on a few main roads, it is more of a pure technology display, and there is no real commercial value.
As a result, AutoX has also been expanding its ODD (Autonomous Driving Design Domain), with a total area of more than 1,000 square kilometers last year. At the same time, Xiao Jianxiong also said that AutoX's RoboTaxi can already cover 95% of the application scenarios, including picking up and dropping off guests at the entrance of the hotel lobby, morning and evening rush hour traffic, and overtaking under complex road conditions.
In addition, Xiao Jianxiong also mentioned the importance of large-scale mass production. The way in which early autonomous driving companies installed hardware on the basis of mass production vehicles was not only inefficient, but also difficult to ensure consistency and reliability. If RoboTaxi wants to achieve true commercialization, it is inevitable that large-scale and standardized assembly capabilities will be indispensable, so AutoX has built the Robotaxi Super Factory to ensure that AutoX's fifth-generation Gen5 RoboTaxi can achieve mass production, which is also the first truly fully unmanned Robotaxi production line in China.
Roundtable Discussion (1): Chip Challenges in the Era of Large Computing Power
Guests: Yang Yuxin, Chief Marketing Officer of Black Sesame Intelligence, Wang Ping, Executive President of Cambrian Xingge, Wang Kai, Director and CEO of Core Qing Technology, And Li Bo, Vice President of Lotus Technology and Head of Intelligent Driving Business Line
2021 is called the first year of lidar on the car, accompanied by the automatic driving computing platform began to exceed 1000 TPOS. This trend in Yang Yuxin's view, the development of automatic driving has come to the "second half of the first half", computing power has also become an important indicator to judge the degree of intelligence of the car, car companies hope to highlight the computing power value, so that end users have more awareness of the automatic driving ability of car companies. The current computing power can theoretically meet the needs of L2+ and L3 autonomous driving systems, and the next focus is to make the scene and experience better.
He also added that "computing power stacking" is a necessary redundancy for subsequent technology upgrades, from the perspective of business logic and technology evolution, chip companies also need to help customers with smaller costs, higher system concentration, lower power consumption, to achieve better automatic driving functions, which is what chip companies have been working hard to promote, but also to promote the evolution of everyone's technology evolution and product routes.
Li Bo, who represents the demand side of the main engine factory, explains the meaning of hardware redundancy from another dimension. He proposed that the hardware defines the software ceiling, reserves enough computing power, reserves enough sensors, and leaves redundancy for the performance requirements of future autonomous driving systems. Otherwise, it's like the current application logic can run through the old phone, but it can't really run.
Wang Ping also mentioned that at present, under the trend of OTA, the car has shown a trend of gradual decoupling of software and hardware. Compared with hardware, the software is easier to upgrade through OTAs, which also prompts car companies to selectively pre-embed in computing power, even if this part is not used now.
In addition, Tesla, Xiaopeng and other car companies self-developed autonomous driving computing chips are also becoming a trend. Wang Kai, director and CEO of Core Qing Technology, said that on the one hand, this is because the shortage of chips makes the main engine factory pay more attention to the diversity of the supply chain and supply security, on the other hand, the high computing power chip has become the core competitiveness of the car company, and the supplier chip is becoming more and more difficult to meet the iteration speed, cost and performance requirements of the main engine factory.
But Wang Kai also said that this route also has many challenges: the threshold of autonomous driving chips is high, and once it takes a detour, it will face huge financial losses, which will also cause incongruity in planning. At the same time, the car grade chip is different from the consumer grade chip, the performance, power consumption and reliability requirements are higher, but also to complete the car grade certification, the cycle is longer, the investment is also larger, the need to recover the upfront cost through the application of a variety of cars, so it is necessary to launch a more inclusive and competitive product system to meet the needs of different car manufacturers.
In addition, the participants also answered the problem of chip shortage. The unanimous view is that the current chip expansion cost is high, and chip manufacturers blindly expand production capacity in the case of not daring to guarantee that there will be the same demand in the next few years. Although the current capacity has recovered from the epidemic, the demand that was suppressed last year has not yet been met, and the real solution may wait until next year.
Roundtable Discussion (2): "How to Mature the Commercialization of Autonomous Driving"
Guests: Zhou Xin, co-founder and chief product officer of Yishi Technology, Hao Jianan, co-founder and chief architect of Tucson Future, Dong Jian, co-founder and VP of software algorithm, and Dai Zhen, vice president of Heduo Technology
Zhou Xin, co-founder and chief product officer of Yishi Technology, and Hao Jianan, co-founder and chief architect of Tucson Future, believe that efficiency and cost are the premises for the commercialization of autonomous driving at the B-end: either to achieve higher efficiency than people, or to achieve fully unmanned autonomous driving. But to achieve the final business logic, not only a very high level of safety and reliability is required, but also the gradual improvement of regulations.
As a company facing both B-end and C-end users, Dong Jian, co-founder and software algorithm VP of Hongjing Intelligent Driving, said that the current landing speed is also faster than imagined, and more mass production models will appear in one or two years. However, due to legal and regulatory issues, most car companies will launch models with L3 level automatic driving experience, but developed according to the L2+ level regulatory system.
Dai Zhen, vice president of Heduo Technology, also gave a more specific time point for the C-end landing of automatic driving - it is expected that 2025 will be a key time node, when the mass production of automatic driving technology, consumer acceptance, infrastructure and laws and regulations will be gradually landed. Before that, autonomous driving in specific scenarios will be the first to land, and Heduo Technology's automatic driving solutions will also begin mass production on a variety of Guangzhou-based vehicle models this year.
Although the "Chief Chi Hang Conference" was held for the first time, it still received strong support from the big names in the industry. In the future, Auto Byte will continue to bring dynamics and analysis in the field of smart mobility, while expanding more projects and contents involving "Zhixing Officials" to present the most cutting-edge and intuitive blockbuster views in the era of intelligence.