The Guangzhou Auto Show, which is due to open at the end of 2022, has aroused widespread enthusiasm in the industry despite being affected by some force majeure. New models, cool technology, luxury equipment… It tried its best to show what an A-class auto show should be. At the auto show, there was a high-profile Youjidu booth. In addition to the “Tianshichengjin” robot finger, it is worth mentioning the appearance of ROBO-01 and the appearance of ROBO-02. ROBO-01 is the first car of Jidu. Based on the empowerment of Baidu AI, it has the sci-fi gene of “car robot”. The concept fully interprets intelligence, and also applies Apollo’s high-level intelligent driving capabilities. It has point-to-point navigation assistance functions when it goes on the market. This car will be delivered this year. How much imagination the “car robot” can bring to users can be glimpsed from the support system of the smart driving and smart cabin behind it.
High-end smart driving ushers in the integration of three domains and starts large-scale mass production
Looking back at this year’s Guangzhou Auto Show, new energy vehicles have occupied a core position in both traditional car companies and new forces. In the first half, we will focus on electric vehicles, and in the second half, we will focus on intelligence. Intelligentization can cover electric vehicles and fuel vehicles, and the intelligentization upgrade battle will appear in a richer form. Taking Jidu as an example, in terms of smart driving, Jidu’s full-stack application of Apollo’s full set of high-level intelligent driving capabilities and safety system has realized the high-level intelligent driving capabilities of “three-domain integration” in the urban domain, high-speed domain, and parking domain. “Three-domain integration” is the capability of Apollo ANP3.0. Apollo ANP3.0 is the industry’s first batch of “three-domain integration” pilot assisted driving solutions that can be mass-produced. At present, Baidu Apollo’s smart driving matrix has also been formed: the AVP parking system will be fully implemented in April 2021; the cost-effective ANP2. This is the only integrated driving and parking solution that can realize parking and high-level functions in high-speed scenarios on a low-computing platform; the ANP3.0 high-level intelligent driving solution has now started generalization tests in many cities such as Beijing, Shanghai, Guangzhou, Shenzhen, and It will be realized this year with the delivery of Jidu. What level can the intelligent driving system of Jidu ROBO-01 reach?In the previous ANP3.0 road test video, the car equipped with ANP3.0 can realize a series of functions such as high-speed and efficient lane change, tunnel passage, large curvature devil curve passage, and autonomous passage through ETC. In the R&D and testing stage, ANP3.0 has already relied on the “BEV Surround View 3D Perception” technology, becoming the only intelligent driving solution in China that can run through multiple scenes in the urban area relying on pure vision. In the road test stage, it adopts pure vision technology, and in the mass production stage, it will adopt a dual-perception and true redundancy solution of “pure vision + lidar” to further cope with unconventional shaped obstacles, night obstacles, and dynamic/static occlusion ghosts from the perspective of perception. Complex urban scenes such as probes and cross-floor parking in parking lots. This is due to Baidu’s L4 technology and L2+ urban smart driving technology, and the algorithm has been verified by more than 40 million kilometers of professional road tests. L4 will continue to provide advanced technology migration and massive data accumulation for L2+ smart driving products, and the feedback of L2 data will also help improve the generalization ability of L4. In addition, based on Baidu’s self-developed navigation map, the city’s high-precision map automatic mapping capabilities, professional map surveying and mapping capabilities ensure the most “familiar road” in China’s road navigation. In terms of data management, ANP3.0 combines Baidu’s deep accumulation in the fields of AI and cloud computing to realize the full lifecycle management of data and the evolution of data-driven smart driving capabilities. Since 2022, the smart driving solution integrating travel and parking has been delivered in batches in mass-produced passenger cars. According to the data of Gaogong Intelligent Vehicle Research Institute, in the first three quarters of 2022, the sales volume of passenger cars equipped with integrated driving and parking in the Chinese market was 498,500, a year-on-year increase of 117.12%. 47.17%. It is predicted that in the next three years, the market space for integrated parking and parking will be about 10 million vehicles. In line with the industry trend, Baidu Apollo delivered ANP2.0, an integrated travel and parking solution, in October 2022, and ANP3.0 will follow up quickly, and will meet consumers this year. As a pioneer in the field of autonomous driving in China, Baidu is also the cradle of autonomous driving talents. Today, the automobile industry has entered the second half of intelligentization. Based on the judgment of the industry, Baidu Apollo has made full preparations for mass production. At present, Baidu Apollo’s smart driving products have been mass-produced on a variety of models such as GAC Aian, Haval Beast, Tank 500, Euler Lightning Cat, and Weimar W6. Baidu Apollo said that in the next five years, the installed capacity of smart driving products will maintain an average annual growth rate of 100%. At the Guangzhou Auto Show, nearly a hundred models equipped with Baidu Apollo smart cabin, smart driving, and smart map, such as Tank 500, GAC Aion V Plus, LX plus, Euler Lightning Cat, and Lynk & Co 09, were all unveiled. It can be seen from this that Baidu Apollo’s intelligent layout is not limited to intelligent driving.
The level of automotive intelligent products varies, and Baidu Apollo “comprehensively” suppresses it
In the big track of smart cars, not only car companies are eagerly occupying positions, AI software and hardware companies in the industry chain have risen one after another, and their smart driving solutions have also emerged with different product combinations. With the scale of products on the car, there are differences in functions and experience. For example, models equipped with ANP2.0 can realize point-to-point automatic driving in high-speed and urban loop scenes; in parking lots, ANP2.0 can not only complete daily parking actions, but also cover complex parking spaces such as mechanical parking spaces. In terms of driving, ANP2.0 can automatically pass through ramps, Y-shaped roads, large curvature curves, tunnels and other road sections, covering more than 30 main driving scenes in total; in terms of parking, it takes 20-30 seconds to complete the parking action, The parking success rate is greater than 95%. This is at the forefront among similar products. For users, product experience is the priority, and for car companies, cost reduction is equally important. How to achieve the experience of high-end models at a low cost, Baidu Apollo gave a different answer. This still has to start with Baidu’s technology accumulation. Baidu Apollo is the only solution in the industry that has a full-stack layout of “cockpit map” and fully mass-produces intelligent solutions for automobiles. In addition to the smart driving mentioned above, in terms of smart cabins, Xiaodu Vehicle has cooperated with more than 70 car brands, covering nearly 600 models and serving more than 5 million users. In terms of smart map, it has car navigation maps and high-precision maps. Among them, the surveying and mapping qualifications of high-precision maps require national approval. A must for a good experience. In addition, Baidu Apollo has also achieved mass production of man-machine co-driving maps, and it is expected that more models will be launched in 2023. Compared with its peers, Baidu Apollo has a more complete business form and a longer business front. There are smart driving, smart cabin, and maps, and these major fields are interactive and interconnected in the process of car intelligence. This will facilitate the sharing of results, and can also be flexibly configured to save costs, providing car companies with Diverse choices. At the same time, the comprehensive business form will help Baidu Apollo obtain more smart car-related certifications. Baidu Apollo has completed more than 10 smart car certifications, which also creates conditions for subsequent mass production and delivery.
Intelligent driving explodes, Tier 1 competes for the finals
2022 is a key node for cars to become intelligent in the second half, and 2023 is bound to intensify competition in terms of mass production scale. Among the companies participating in this track, there are traditional Tier 1 companies, such as Bosch, which has been designated by OEMs for L2+ intelligent driving products, and plans to launch at the end of 2023; there are emerging L4 autonomous driving companies, such as Qingzhou Zhihang , Momenta, etc., are rapidly launching high-end smart driving products to seek orders; there are also self-developed car companies, such as the new forces, who pursue new technologies and hold core technologies in their own hands as much as possible. However, when the automotive intelligent industry chain matures, from the perspective of commercial logic, the industry needs several excellent suppliers to carry the banner.By 2030, the penetration rate of advanced intelligent driving will increase to over 90%. So in the next few years, who will be the representative of smart car solution providers? Who will be the intelligent Tier1 leader? In the era of intelligence, one of the basic characteristics of Tier1 is specialization. With the increase in the number of mass-produced vehicles, intelligent driving systems and other intelligent products are widely spread, so a strict quality certification system is a major prerequisite. The second is new technology. High-level intelligent driving requires powerful algorithms and software capabilities, as well as the integration capabilities of vehicle regulation systems.This is the reason why traditional Tier1 iteration speed is not as fast as new algorithm companies. The third is localization. The road environment, laws and regulations, culture, and habits of each region are different. As China is the main front for the development of intelligent automobiles, product design and communication efficiency for localization will be very critical.
Baidu Apollo is committed to becoming a Tier 1 that serves OEMs in the era of automotive intelligence. The snowball technology accumulation effect brought about by its early investment in autonomous driving has begun to appear. Baidu’s autonomous driving technology experts have said that Baidu’s confidence and confidence in empowering cars with intelligence stems from the accumulation of autonomous driving technology in the past ten years. Then, the technological evolution and product experience of car intelligence are the bigger test. On the track of intelligent cars represented by intelligent driving, there are still many possibilities for Tier 1, and there are still variables to determine who will win in the end. However, continuous R&D investment and the pace of mass production are the two sharp edges of Baidu Apollo.
Comments