(**Gasgoo).
On December 14, BMW Group's vehicles equipped with Level 3 autonomous driving functions officially obtained a high-speed autonomous driving test license in Shanghai.
BMW said that it is fully prepared for the adaptation and application of L3 autonomous driving functions in China in the future, so as to meet the complex traffic conditions in China and the driving characteristics of Chinese customers. We will pay close attention to the requirements of Chinese regulations and bring Level 3 autonomous driving capabilities to the market at an appropriate time.
Gasgoo noticed that BMW's official website recently disclosed a document of the "BMW China Summit", which is the core partner of its intelligent driving China solution. Well-known technology companies at home and abroad, such as Qualcomm, Intel, Valeo, Tencent, Qianxun, NavInfo, etc., are among BMW's leading companies in intelligent driving. Among them, the "high-precision positioning" solution provided by Qianxun Location + NavInfo spans two generations of platforms.
Not only BMW, but also mainstream Chinese auto brands such as NIO, Xpeng, Li and GAC Aion have gradually equipped their new cars with high-precision positioning systems as standard.
Benefiting from the dual drive of industry and policy, the development of high-level autonomous driving above L3 is accelerating around the world. With the L2+ to L3 transition, the decision-making and control of the vehicle will be "completely" done by the system under certain conditions. The demand for high-precision positioning solutions will be higher and higher.
According to the understanding of Gasgoo, at present, the high-precision positioning scheme is becoming more and more mature, and it has dropped to the level of market acceptance, and the penetration rate of the terminal market continues to increase. According to data from the Gasgoo Automotive Research Institute, in the first 10 months of 2023, the penetration rate of high-precision positioning systems for new cars in China has increased from 121% is gradually approaching 5%, which is between the small-scale introduction period and the large-scale expansion (bounded by 10% penetration).
BMW China's choice
China is not only BMW's largest single market, but also an important R&D base. At present, BMW has built a local R&D team of 3,200 people in China, and has launched many locally designed products, such as navigation, WeChat, etc. designed for the Chinese market.
For a long time, BMW's R&D team in China has been carrying out localized research and development of Level 3 autonomous driving. As a long-established German car company, BMW's intelligent driving solutions usually release signals for the industry.
BMW).
According to the information disclosed on BMW's official website, BMW's next-generation intelligent driving technology platform Gen AD25 in China has adopted cooperation with Qualcomm, Valeo, Qianxun Location, NavInfo and others. The original Gen AD21 technology platform is in cooperation with Intel, Mobileye, Qianxun, NavInfo, Stellantis, Tencent, etc.
Among the two generations of intelligent driving technology platforms, Qianxun, China's leading high-precision positioning company, and NavInfo, a high-precision map company, span two generations of technical solutions - it can be seen that BMW, which attaches great importance to safety, firmly chooses the "high-precision positioning + map" solution.
BMW actually started the development of the L3 very early, around 2018. BMW is looking for local technology partners in China. Industry insiders revealed.
BMW).
BMW did not provide more information about the L3 intelligent driving solution, and neither Qianxun Location nor NavInfo commented on the details of the cooperation.
According to public information, Qianxun Location, the partner of the first company, is responsible for the construction and operation of the national Beidou ground-based augmentation system, which was established in 2015 to provide users with centimeter-level positioning, millimeter-level perception, and nanosecond-level spatio-temporal intelligent services. Qianxun has built more than 4,500 GNSS satellite-based ground-based augmentation stations all over the world, self-developed positioning algorithms and large-scale Internet service platforms. It covers five major areas: intelligent driving, public services, industry upgrading, smart cities, and mass consumption.
Taking the cooperation between Qianxun and Xpeng Motors as an example, with the help of the high-precision space-time service network of Qianxun Location, the global positioning accuracy of Xpeng Motors P7 intelligent coupe can reach the centimeter level, and NGP high-speed autonomous navigation and driving can be realized on the highways covered by high-precision maps. Xpeng Motors has said that in bad weather, non-line-of-sight scenarios or other scenarios that have a certain impact on the work of on-board sensors, Qianxun's location can still provide absolute positioning results for the car and provide a decisive reference for the car's direction perception.
Another partner, NavInfo, as a listed company, announced its agreement with BMW (China) in February this year. NavInfo will provide products and services such as ADAS (Advanced Driver Assistance System) maps, HD maps and LBS (location-based services) for the BMW Group's next-generation autonomous driving functions in the Chinese market. The announcement revealed multiple levels of cooperation between the two parties, including the construction and service of IoV-related platforms such as in-vehicle navigation maps, IoV location data platforms, and CNS vehicle interconnection systems. The agreement is valid until December 2033.
German car companies have always attached great importance to safety, so safety has also become a very important indicator for them to design intelligent driving systems. In this context, the high-precision positioning scheme has its own unique value. An industry insider revealed to Gasgoo.
From L2 to L3, the demand for high-precision positioning has increased
According to the analysis of industry insiders, the "high-precision positioning + map" scheme has become the core foundation to ensure the overall safety and reliability of the intelligent driving system, and can support different application scenarios and hardware architectures of L2-L3 autonomous driving systems.
The key technologies of intelligent driving are perception, planning, decision-making, and execution. According to the operating logic of self-driving cars, the premise of the vehicle to complete the autonomous driving from point A to point B is to determine "where I am", and then solve the next "where to go" and "how to go" problems.
In the L1 and L2 stages, since the driver is the main environmental information perception and processor, the positioning accuracy of the vehicle can basically meet the needs at the meter level.
However, for L3, because the control of the vehicle can be handed over to the system in some cases, in special scenarios such as urban canyons, viaducts, tunnels, underground garages, etc., the vehicle's positioning accuracy requirements must reach the centimeter level, and have sufficient high availability to achieve the highest possible driving safety.
In particular, with the transfer of applications to urban scenarios, the disadvantages of the perception scheme that relied on high-precision maps for intelligent driving in the past have been continuously highlighted, and they have been replaced by the "heavy perception + light map" scheme, which is a key part of vehicle-end perception, and the importance of high-precision positioning has become more prominent. In the new architecture of "BEV + Transformer + Data Closed-Loop", high-precision positioning provides underlying support in the system matching and using ADAS maps and improving the effect of BEV algorithms (so as to have the ability of vehicle-side SLAM mapping and the ability to perceive the true value of samples annotated).
Satellite-based high-precision positioning is the only sensor in the vehicle that can provide absolute position and accurate time, which can assist relative positioning technologies such as vision, radar, and inertial navigation to match environmental features. This determines the natural irreplaceability of satellite high-precision positioning in the intelligent driving system.
Driven by multiple factors such as policy, technology and scenario demand, the inflection point of intelligent driving is gradually emerging. With the advent of L3 high-end intelligent driving, the value of the industrial chain has ushered in new opportunities, which is expected to further promote high-precision positioning and lidar on the car and play a real role.
In terms of policy, from the "Guidelines for Transportation Safety Services for Autonomous Vehicles (Trial)" issued by the Ministry of Communications on December 5 to the "Notice on Carrying out the Pilot Work on the Access and Road Access of Intelligent Connected Vehicles" issued by the Ministry of Industry and Information Technology, the Ministry of Public Security, the Ministry of Housing and Urban-Rural Development, and the Ministry of Communications on November 17, the L3 policy has been intensively implemented recently.
Technically, the battlefield of intelligent driving warfare has shifted from closed highways to more complex urban scenarios. Urban pilots have been landed one after another. In 2023, almost all the leading players participating in the smart car competition in China will release the NOA pilot assistance function in urban areas, that is, it supports point-to-point assisted driving according to the navigation route under urban roads.
At present, whether it is bicycle intelligence or vehicle-road coordination, high-precision positioning will play an increasingly important role.
Shenwan Hongyuan's research pointed out that whether from the perspective of product maturity and mass application cost, or from the perspective of the willingness of terminal car companies, high-precision positioning will be widely configured as the hardware redundancy of the high-level autonomous driving perception layer, and will become the standard configuration of L3 and autonomous vehicles.
The "high-precision positioning" scheme has entered the era of mass production
As more and more car companies focus on high-end intelligent driving, the high-precision positioning track is becoming more and more lively.
At present, the participants in the high-precision positioning industry chain include chip (Huada Beidou), satellite navigation, inertial navigation and other component manufacturers (Daoyuan Electronics, Juefei Technology, etc.), positioning service providers (Liufen Technology, China Mobile, Dayouspace, etc.), integrated operation service providers (Qianxun location, etc.), Tier1 (Bosch, Continental, etc.), etc.
According to industry insiders, the cost of high-precision positioning services + terminals has dropped from 10,000 yuan three or four years ago to 1,000 yuan, and has basically dropped to the mass production range recognized by the automobile market.
After several years of continuous construction, the integration of satellite-based ground-based augmentation system has been completed, as well as the chip integration of high-precision positioning terminals, and the technical scheme of high-precision navigation equipment has been significantly simplified, and obvious scale advantages have been formed. The source revealed.
Taking the mass production location as an example, it can be used by car companies to complete the evaluation within 1 month, complete the adaptation within 3 months, and achieve mass production within 6 months. This solution is summarized by the company as "automotive high-precision positioning industrialization system". According to the latest information released by Xinhuanet, this set of solutions has cooperated with more than 40 car brands, and has been delivered in batches on more than 30 models, with a cumulative ** service of more than 1 billion hours.
Industry standards are also in full swing. Recently, led by the Highway Research Institute of the Ministry of Transport, Qianxun Location, AutoNavi, Huawei, etc. jointly compiled the "Technical Guidelines for Highway Engineering Facilities to Support Autonomous Driving" to unify a number of infrastructure and technical indicators.
It is obvious that the high-precision positioning solution has entered the era of mass production. Driven by leading manufacturers, the cost performance, delivery speed, service standards (reliability, integrity, etc.), positioning algorithms, and operation and maintenance quality have all been greatly improved compared with the past few years.
Intelligent driving technology is a system engineering, just like a wooden barrel, and there can be no shortcomings. The above-mentioned people believe that "the car is equipped with more and more sensors, high-precision positioning, cameras, millimeter-wave radar, lidar, etc., but how to make full use of the capabilities of these sensors is the key, not just stacking, to think holistically to convert these capabilities into system capabilities." In addition, the data resources, computing architecture, and algorithm capabilities of the entire intelligent vehicle also need to be considered as a whole. ”