With the advent of the "new four modernizations" wave, traditional car companies are transforming to electrification and intelligence, and the first batch of new car-making forces have ushered in a counterattack, not only Tesla's market value has exceeded one trillion US dollars, Wei Xiaoli's market value has also exceeded 10 billion. In the context of a lively scene, where are the opportunities for newly entered Internet car manufacturers?
As a newly entered representative of Internet car manufacturing, Xia Yiping, CEO of Jidu Automobile, was interviewed by AutoMotive Observation during the China Electric Vehicle 100 Forum (2022) on the occasion of the first anniversary of the establishment of Jidu Automobile, responding to its thinking on the entire industry and the latest progress of Jidu.
The car building won the second half with AI
The industry generally believes that the first half of car manufacturing is new energy, and the second half is intelligent networking. Today, the first half has been successful and the second half has begun. Xia Yiping believes that the advantages of enterprises in AI intelligence are the core factors for winning in the second half of car manufacturing.
Xia Yiping pointed out that in the era of fuel vehicles and electric vehicles, the innovation of the automobile industry is driven by energy. Today, electric vehicles are ushering in a comprehensive digital upgrade, and AI intelligence has become a new trend. "AI has brought about technological innovation, efficiency improvement, and experience disruption. We believe that 2023 will be the first year of automotive intelligent competition, and the era of intelligent car 3.0 has arrived. ”
Throughout the industry, for a long time in the past, the chip computing power of cars has been lower than that of smart phones, which has led to many AI technologies that cannot play an advantage in cars and cannot comprehensively improve the user experience. Nowadays, the emergence of high-computing car chips such as Qualcomm 8295, Nvidia, and Horizon has gradually transformed cars from transportation tools into ai-driven intelligent mobile spaces.
In Xia Yiping's view, the products of the smart car 3.0 era should have the following three characteristics:
First, with hardware and software systems that support high-level automatic driving capabilities, the autonomous driving of automobiles has grown from only responding to specific scenarios to calmly coping with more complex scenarios.
Second, speech semantic interaction should be accurately recognized, and the communication process should be natural.
Third, according to the user's habits of using the car, complete the independent learning and functional iteration, so that the car understands the user more and more.
Behind these experiences and capabilities, there is actually a complete set of complex software and algorithms, as well as hardware that meets the corresponding computing power requirements, which can be realized.
Xia Yiping believes that in the era of smart car 3.0, whether it is automatic driving or the AI capabilities of the intelligent cockpit, it puts forward high requirements for the computing power in the car, and the redundancy of the computing power improves the certainty of the landing of high-level automatic driving capabilities. At the same time, in order to achieve the integrity and stability of high-order autonomous driving algorithms, including high-definition cameras, millimeter-wave radar, lidar, etc. will become the standard hardware of high-order autonomous driving intelligent cars. In addition, configurations such as domain controllers and smart chassis for high-end autonomous driving will also become standard for smart cars.
The safety and stability of smart driving and smart cabin software operation, in addition to relying on the protection of high-level hardware, but also in the ability of software algorithms, and through the safe mileage of data accumulation, to make car robots smarter. "AI needs to integrate hard and soft, hardware security comes from quality, system and appropriate computing power redundancy; software security comes from the accumulation of big data in a wider range of scenarios and faster iteration of algorithms." Xia Yiping stressed.
Therefore, the research and development of software, the verification of software capabilities, and the safety and stability of software operation have become the core of jidu automobiles in the intelligent car 3.0 era.
The advantage of intelligent first-mover is highlighted
Compared with this wave of new car-making forces such as Wei Xiaoli, Jidu Automobile, which was jointly established in March 2021 by Baidu and Geely, is a proper latecomer. Losing the first-mover advantage in time, what does Jidu Automobile rely on to participate in the industry competition?
It is reported that as a subsidiary of Baidu, Jidu Automobile directly inherits Baidu Apollo intelligent driving technology and Robotaxi's huge data accumulation.
As we all know, Baidu is the earliest company in China to lay out autonomous driving, and its Apollo's Robotaxi has high-end automatic driving capabilities, and the technology is in the first echelon of the world. In 2021, Baidu Robotaxi took the lead in launching in Beijing Shougang Park, and has now entered the social open road in the core areas of Beijing Yizhuang and other core areas, and the commercial toll operation has also been applied to the daily travel services of the public from the park. The autonomous driving test mileage is accelerating at a rate of no less than 1 million km/month. At present, Baidu Apollo has accumulated more than 25 million kilometers of safety test mileage, and continues to maintain safe operation with zero accidents, which is unmatched by other start-ups.
Because of this, Xia Yiping confidently told reporters that the competitive advantage of Jidu Automobile is mainly in the ability of intelligence. "This ability does not refer to the intelligent ability of the past, in the era of intelligent cars, intelligent capabilities are actually driven by AI." Jidu is the only smart car brand in the industry that applies Baidu Apollo's high-end autonomous driving capabilities in the industry. ”
Xia Yiping further explained: "Intelligence is actually a beginning, the core of driving intelligence is that high-end high-computing chips have begun to be applied on a large scale in the car, and they are widely accepted by various car companies. In the past, some artificial intelligence algorithms, especially Baidu, have been practicing for a long time, and the real high-level autonomous driving algorithms need to run in the cloud; the most core change brought by high-computing chips is to let these algorithms land in the car. The competitive advantage of Jidu Automobile is reflected in how to create benchmark products through AI-driven product intelligence and experience innovation in the future. ”
Xia Yiping frankly said that one of the tasks of Jidu Automobile is to productize the autonomous driving capabilities accumulated by Baidu for more than a decade, and push highly intelligent products to users.
Although all the core capabilities of Jidu Automobile currently come from Baidu Apollo autonomous driving software, Xia Yiping stressed that Jidu's application of Baidu Apollo's high-level automatic driving ability is actually different from Robotaxi, and the problems solved by the two are not the same: Robotaxi solves the problem of early verification and commercialization of unmanned autonomous driving capabilities; and Jidu is for mass production and consumers, with the goal of high-end autonomous driving product experience. Re-integration and verification development of Baidu Apollo's high-level autonomous driving capabilities.
Thanks to the accumulation of tens of millions of kilometers of road test data in Robotaxi, the software development process of Jidu automotive robots can be maintained efficiently.
To this end, Jidu Automobile has developed a set of research and development processes exclusive to the "brain" of Jidu Automobile Robot - SIMUCar (Software Integration Simulation Sample Car), which realizes the decoupling of the research and development of the vehicle's software and hardware, without interfering with each other, dual-track parallel, and pre-positioning the development of high-level automatic driving capabilities, so as to exchange more time for the "safety and stability verification" of the software.
It is reported that in September 2021, SIMUCar has started running; in November 2021, SIMUCar launched a dynamic test, synchronously realizing the development of intelligent driving capabilities and the development of smart phones updated once a week, successfully verifying the reliability of functions such as signal light recognition, right turn at intersections, and driving at night; in December 2021, SIMUCar updated the application of autonomous driving scenarios common in urban and high-speed areas such as left turn at intersections and autonomous lane changes; at the end of February 2022, SIMUCar was upgraded to version 2.0, and at this stage, the integration of high-level autonomous driving software, vehicle electronic and electrical architecture and SOA has been completed.
According to Xia Yiping, SIMUCar will iterate in "weeks" in the future. It is expected that by the end of the year, intelligent assisted driving in the urban and high-speed areas will be realized. "Jidu's autonomous driving iteration capability is very fast, and before the product is launched, users can experience the automatic driving function during test drives, not after the listing, through OTA upgrades."
Mass production of the first automotive robot in 2023
Looking at the vast majority of car companies in the industry, a car from research and development to mass production delivery, generally takes 3-5 years, but Jidu Automobile has set this time at 2 years. At present, the time has passed the halfway point, what stage has the progress of Jidu Automobile's car manufacturing reached?
According to Xia Yiping, the current production and research and development progress of Jidu Automobile is very fast, January and February 2022 is the first car mass production finalization, and suppliers began to manufacture molds in an important stage, is currently preparing some production verification process A sample, B sample, C sample.
It is worth mentioning that in order to ensure that the recent epidemic does not affect the progress of research and development, the team took the initiative to work in the company during the closure period. From the current situation, the overall pace of research and development has not been delayed.
When asked whether the chip shortage will affect the mass production schedule of Jidu Automobile, Xia Yiping frankly said: "Although we have not yet mass-produced and put into production, in the production verification stage of the prototype, there will also be some chips that cannot be found." If there is no way to find a chip that supports product validation, the entire development cycle will be extended. To ensure that the schedule is not affected by the lack of cores, we have now set up a supply team. Generally speaking, the warranty team was not established until the eve of mass production, but now it is established at the stage of testing prototypes. In addition, the tight supply of chips will also bring about a great increase in costs, and sometimes the cost of a sample is close to dozens of times the cost of mass production, which is very exaggerated. Sometimes we have to take cash to buy spot, we look for chips every day in the country or globally, it is two or three to find, not in batches. ”
Xia Yiping firmly believes that the overall supply chain shortage environment will be alleviated in the second half of 2022. Judging from the production time of Jidu automobiles, the supply chain will not have a greater impact on mass production.
The research and development of the car is only the tip of the iceberg, and there are many challenges next, such as channels, sales, production, supply chain, production and marketing coordination, raw materials, etc. However, Xia Yiping believes that these are the only ways for startups to grow, and Jidu Automobile will face all challenges.
According to the plan, Jidu Automobile will release the first concept car of automotive robot at the 2022 Beijing Auto Show, and its first mass-produced automotive robot will be delivered in 2023, which will have high-end automatic driving capabilities.
In terms of channels, Xia Yiping introduced that Jidu Automobile will adopt the model of self-built channels, and the first flagship store this year will land in Shanghai.
Although it will take time for the first product to be mass-produced, Xia Yiping firmly believes that the market will give a positive response to Jidu products, whether it is the design of internal and external decorations, functional definition, or product experience. "During the development process, we continued to find a lot of people, through the external perspective of the car, no matter what field of people came to visit, the feedback is very good."
While full of confidence in the first product, Jidu Automobile has also opened the pre-research of subsequent models. It is reported that the second model will be unveiled at this year's Guangzhou Auto Show.