At the Guangzhou Auto Show just past, autonomous driving technology has once again become the focus of attention. At this auto show, the autonomous driving technology exhibition hall and the autonomous driving experience zone were introduced for the first time. Passenger cars including the Great Wall Salon Mechalong, Weimar M7, Xiaopeng G9, etc. have also begun to carry multiple lidars. Show your own level of autonomous driving.
Corresponding to the successive increase in the number of lidars by independent brands is Tesla’s preference for pure visual solutions. In the field of autonomous driving, whether the pure vision solution will dominate the future, or the vision + lidar hybrid technology will go further, discussions on this topic have never stopped. In this regard, Juefei Technology CEO Li Dongmin pointed out that for market demand, it is necessary to take the route of multi-sensor fusion.
The pure visual solution is yet to be discussed
The pure vision solution is a solution dominated by a camera. The vehicle builds a neural network through artificial intelligence and deep learning. However, because the camera is less accurate, this solution has higher requirements for algorithms and computing capabilities. Compared with lidar, low cost becomes the biggest advantage of this scheme. But for this technology, Juefei Technology has maintained a clear-cut attitude of opposition, and “safety” has become the main reason for Juefei Technology’s opposition.
“The pure vision solution may be able to achieve high-level automatic driving, but the cost is relatively high.” Li Dongmin analyzed that from a technical point of view, automatic driving should pay attention to the level of misoperation and accuracy, but in some extreme cases , The control of the vehicle by the pure vision solution will still be problematic.
“For a typical example, when there is a road maintenance vehicle or an overturned truck parked in the middle of the road, and these vehicles happen to be silver or white paint, then the pure visual solution will have a high probability of problems.” Li Dongmin said . Regardless of the age, ensuring safety is the bottom line of automotive products. The shortcomings of the pure vision solution in terms of safety and security have caused most car companies except Tesla to choose a multi-sensor fusion autonomous driving solution.
The multi-sensor fusion scheme refers to the automatic driving technology that integrates multiple schemes such as camera, lidar, millimeter wave radar and so on. The multi-sensor fusion scheme can give full play to the strengths of various technologies to ensure the safety of automatic driving of vehicles with redundant technologies. While fully supporting the multi-sensor fusion technology, Li Dongmin also pointed out that this kind of technical solution needs to be continuously improved in the two areas of vehicle-end and vehicle-road collaboration in the future.
In terms of vehicle end, lidar technology still needs continuous breakthroughs. At present, the longer the life of the lidar, the worse the detection range, which requires a balance between the life of the lidar and the detection range.
In terms of vehicle-road collaboration, the roadside and vehicle-side data also need to complement each other. In Li Dongmin’s view, the relationship between cars and roads is like the relationship between car lights and street lights. Not all roads have street lights. For example, straight roads such as expressways do not need street lights, while entering tunnels, ETC toll stations, and ramp junctions require “street lights”.
“The role of the roadside is to provide the driver with a beyond visual range, especially the blind spot information to the car,” Li Dongmin analyzed, “No matter how the sensors on the car develops, the blind spot problem cannot be solved. Collaborative matching can achieve this effect.”
Layout the passenger car autonomous driving market
Based on digital roads and multi-sensor fusion as its core technology, Jufei Technology, which provides full-scenario navigation services for autonomous vehicles, has keenly captured market opportunities and has become a leader in this field. In November, Juefei Technology won the “2021 First Batch of High-tech Enterprises” issued by the Ministry of Science and Technology and the “Sixth Batch of Specialized Professionals” issued by Beijing Municipal Bureau of “Jingtexin SMEs” dual certification.
In terms of partners, from the leading commercial OEMs to the chip giants, to the in-depth participants in the traffic junction car-road collaboration industry, all have cooperated with Jufei Technology. “The self-driving bus that Juefei cooperated with Yutong Bus and equipped with a fusion positioning AR system has landed in Zhengzhou, Henan Province last year. We have maintained a very good cooperative relationship until now.” Li Dongmin said.
Juefei Technology’s current achievements are directly related to its keen market sense and advanced technology development level. Recalling the experience of developing the commercial vehicle market, Li Dongmin said: “Before the autonomous driving market has developed, we built a data engine. This engine uses data and multi-sensor fusion as its core technology to empower autonomous vehicles in this way. , To provide full-stack integrated computing solutions for autonomous driving, including integrated perception and integrated positioning.” With the booming L4 commercial vehicle market, commercial vehicle OEMs began to look for large-scale, commercialized operations in the market. The solution provider, Juefei Technology, which has outstanding comprehensive strength, has been assigned the biggest cake smoothly.
Juefei Technology Multi-sensor Fusion Perceives Holographic Intersections
At present, Juefei Technology’s services mainly include two categories, one is car OEM OEMs and first- and second-level component suppliers, and the other is autonomous driving operators. Autonomous driving operators refer to companies that do not take car building as their core purpose, but need to use vehicles to complete commercial operations.
As we all know, the research and development of autonomous driving technology is quite expensive. In terms of financing, Juefei Technology has always maintained its own clear rhythm. At present, Jufei Technology has completed the A+ round of financing. These funds are mainly used for the research and development of multi-sensor fusion AI technology and autonomous driving solutions. In the next B round of financing, Jufei Technology will target the field of passenger car autonomous driving.
“2020-2030 will be the 10 years for the golden development of autonomous driving and vehicle-road synergy. The industry will complete the large-scale application of autonomous driving from commercial vehicle scenarios to passenger vehicles.” Li Dongmin believes that by 2030, passenger vehicles The number of smart cars will be greatly increased, and the data and technical capabilities accumulated by Jufei Technology will become the key support for the coordinated development of autonomous driving and roads.
To this end, Jufei Technology will accelerate the deployment of autonomous driving technology in passenger cars. Juefei Technology believes that the establishment of fusion perception capabilities and data closed-loop capabilities based on mass-produced car lidar is a key reliance for the realization of passenger car NOA systems. In order to build fusion perception and data closed-loop capabilities, Jufei Technology has deployed an L4-level multi-sensor fusion perception system on the roadside on a large scale to efficiently collect all kinds of traffic data from a global perspective and build a training network model around the clock. Such an approach not only creates a technological opportunity for vehicle-road collaboration, but also lays a solid data foundation for the “carriage” of the fusion perception system.
2022 is considered to be the first year of mass production of L3 passenger cars (or high-level assisted driving). Facing the upcoming challenges, I feel that non-technologies are eager to try and are full of confidence. Li Dongmin emphasized: “In the field of commercial vehicle autonomous driving, we have already obtained the approval of the OEM, and this trust will be the best help for us to impact the passenger car market.”