DJI Introduces High-End Mass-Produced Intelligent Vehicles

On September 26th, the Baojun Yunduo 460 Pro and 460 Max Lingxi Edition models were officially launched and delivered. The Baojun Yunduo 460 Pro Lingxi Edition is priced at $17,200, while the 460 Max Lingxi Edition is priced at $18,300.

The Baojun Yunduo’s Lingxi Edition is equipped with the “Lingxi Intelligent Driving 2.0” system developed in collaboration between SAIC-GM-Wuling and DJI Automotive. This system’s capabilities are supported by DJI Automotive’s “Chengxing Platform” and it is also the first mass-produced result of the basic version of Chengxing Platform.

On July 27th, DJI Automotive announced that its new generation of the intelligent driving solution is named “Cheng Xing”. This platform uses inertial navigation and stereo binoculars as the core perception components. It has two pure visual configurations: 7V and 9V, and supports various sensors such as expanded LiDAR. It can achieve multiple L2+ advanced intelligent driving functions on a configuration with computing power as low as 32 TOPS.

The Baojun Yunduo’s two Lingxi Editions are both equipped with the 7V (1 inertial navigation dual camera + 5 cameras) configuration on the production platform. They are equipped with a pair of 8 million pixel inertial navigation dual cameras, four 3 million pixel surround view cameras, one 3 million pixel rear monocular camera, as well as an additional forward millimeter-wave radar and twelve ultrasonic radars.

According to official information, these two models can already achieve intelligent driving assistance in the speed range of 0-130 km/h, high-speed navigation assistance without relying on high-precision maps (Navigate on Pilot – NOP), long-range memory parking up to 1000 meters, and advanced intelligent driving functions such as 100-meter reverse tracking. Additionally, a memory navigation assistance optional package will be released later this year, which supports city navigation assistance for up to 10 routes longer than 100 kilometers.

In April 2021, DJI officially launched its intelligent driving business brand “DJI Automotive” and announced that it would become a Tier 1 supplier of autonomous driving solutions. Currently, the car manufacturers that DJI has publicly announced partnerships with include Volkswagen and Wuling. In June of last year, the first mass-produced vehicle equipped with DJI Automotive’s intelligent driving system was released – SAIC-GM-Wuling’s 2023 KiWi EV.

SEE ALSO: DJI Updates Mavic 3 Drones with Enhanced Features

This year, DJI has launched a new generation of an intelligent driving solution called the “Chengxing Platform”. The biggest feature of this platform is its pure visual solution, which does not rely on high-precision maps and expensive components such as LiDAR. However, it can also support expansion if needed.

The technological core of the Chengxing platform is the “inertial navigation stereo binocular vision perception system”. It obtains distance information and generates dense point clouds using visual principles similar to human eyes, solving the recognition and bypassing of irregular obstacles. It can also help vehicles predict blind spots ahead and prevent potential visual obstructions. Like human eyes, it sees the three-dimensional world, but with slightly less accuracy compared to lidar.

This technology comes from DJI’s accumulation in the drone industry. The visual recognition technology involved in drones happens to be a key technology for autonomous driving development as well. Apart from differences in safety requirements, there are many similarities between the two in perception, decision-making, planning, and execution.

There has always been a debate in the field of autonomous driving between two approaches: the pure vision route or the fusion of vision and LiDAR. Tesla represents the former, as it has achieved lower hardware costs compared to its competitors due to its powerful algorithm advantage. However, this approach requires extremely high demands on data and algorithms, making it a threshold that many car manufacturers find difficult to overcome.

DJI’s idea is to build a platform with strong scalability, which can achieve pure vision and also flexibly adapt to the development needs of different vehicle models, reducing the cycle and cost of mass production for advanced autonomous driving.

Some industry professionals also believe that the dual-camera inertial navigation system has a high technical threshold, especially in mass production, which requires significant demands on installation processes and long-term stability. Implementing it in automobiles is even more challenging than in drones. This is also a challenge that DJI needs to face.