laitimes

Joint venture learning new forces, Nissan's new self-driving program is "imitation" of Weilai?

Joint venture learning new forces, Nissan's new self-driving program is "imitation" of Weilai?

The new energy models of domestic new forces to build cars, in terms of intelligence, can almost crush the current joint venture models, especially in terms of intelligent driving, the joint venture brand can cover the basic L2 level auxiliary driving, intelligent driving perception hardware is also quite basic, for the pursuit of high-level intelligent driving, it seems that it is not as urgent as the new forces to build cars. However, at a time when cars are making great strides towards electrification and intelligence, joint venture model brands have also begun to gradually exert efforts in intelligent driving.

Joint venture learning new forces, Nissan's new self-driving program is "imitation" of Weilai?

Nissan recently announced the latest development progress of the new Driver Assist ProPilot, which was previously introduced with the ProPilot 1.0 Driver Assistance System, which is currently equipped with this function on models such as Tianlai and Qijun, which realizes the ACC adaptive cruise function through sensors such as the front-mounted monocular camera and millimeter wave radar, and also has a certain lane-keeping capability, but the driver must put his hand on the steering wheel when using this function. This system is a very conventional level in the assisted driving system of the joint venture brand, and there is still a big gap between the assisted driving of the new forces and the car manufacturers.

Joint venture learning new forces, Nissan's new self-driving program is "imitation" of Weilai?

The Nissan plan 3 years ago is no different from Wei Xiaoli?

The ProPilot 2.0 system, which is currently used in Japan, is a system that Nissan hopes to compete with new forces such as Tesla. This 2019 system and the new forces at that time made cars, which are not much different in the hardware part of intelligent driving perception, belong to the mainstream level of 2019: it has a total of 8 cameras (3 forward vision + 4 surround view + 1 car), 5 millimeter wave radar (1 forward + 4 angle radar), 12 ultrasonic radars, plus high-precision maps.

This set of perception hardware systems and the Weilai ES8 of that year are almost exactly the same. Relying on this sensor configuration, Nissan can realize high-speed navigation assistance driving function, similar to the high-speed NGP of the new domestic forces, which have delivered high-speed navigation assistance. This system can also achieve off-the-hand driving in a single lane, this function has not really "started" in the domestic new forces, but under the existing hardware conditions, the experimental demonstration has been relatively sufficient, Xiaopeng's next stage of the urban NGP is aimed at this function, the current overseas Tesla models have been realized in some areas to achieve this function, in the mainland also need to wait for further follow-up and improvement of laws and regulations.

Joint venture learning new forces, Nissan's new self-driving program is "imitation" of Weilai?

In fact, from the above functions and hardware configurations, it is not difficult for us to see that the system functions of Nissan and the system functions of the main sales models of the new domestic forces are actually relatively similar, and there is no gap, but this is after all the level shown by the basic hardware 3 years ago, and now in this intelligent "next stage", let's take a look at the difference between traditional car-making and new forces in terms of intelligent driving.

Nissan plans to complete the development of "road condition perception" technology in the mid-2020s and apply it to all new models by 2030. The biggest difference between road reality perception technology and the previous one is the addition of lidar, which is displayed on real vehicles, that is, ProPilot 3.0.

Joint venture learning new forces, Nissan's new self-driving program is "imitation" of Weilai?

Is Nissan's new system mimicking NT2.0's NIO?

Nissan Motor Announced a Partnership with Luminar to integrate next-generation lidar into Nissan's driver assistance system, and the new system will be further enriched in terms of intelligent driving hardware: 10 cameras (1 front camera for detecting the front view, 2 front view cameras on the left and right sides of the roof, 2 rear view cameras on the left and right sides of the fenders, 1 rear view arranged at the shark fins on the roof to identify the rear environment, these 6 cameras form a 360-degree field of view, The other 4 belong to the fisheye camera, one in the front and one in the front and one in the rearview mirror on the left and right should be used for parking). 7 millimeter-wave radars (4 short-range radars arranged in 4 corners of the vehicle, 1 long-range radar in front and 2 long-range radars in the back), 1 luminar lidar (arranged on the front roof), 1 high-precision map module. There is no mention of ultrasonic sensors, which are more traditional and inexpensive sensors and should be equipped. This set of hardware perception system and the new flagship model of the new force Wei Xiaoli, in terms of intelligent driving hardware, the configuration is more similar, and it is a planning arrangement made in order to achieve L3-L4 level automatic driving.

Joint venture learning new forces, Nissan's new self-driving program is "imitation" of Weilai?

In terms of layout, Nissan's system can be regarded as a layout that we are particularly familiar with. The lidar was placed on the roof of the car, which can be seen as an endorsement of Li Xiang's proposition... At the same time, the side view camera is also arranged on the left and right sides of the model, and the layout of the entire roof sensor is similar to that of NIO ET5 and ET7, while NIO is regarded as a monocular and Nissan is binocular. There are also two forward-facing and backward-facing cameras at the rearview mirrors of the side fenders of the test vehicle, which can be seen in the perception of the front of the side, and the overlapping and redundancy is carried out by two sets of cameras and millimeter-wave radar.

Joint venture learning new forces, Nissan's new self-driving program is "imitation" of Weilai?

And Nissan mentioned that it will use Luminar's next-generation lidar, which we will mainly consider three parameters: the detection angle of the FOV field of view, including horizontal and vertical angles, the maximum detection distance, and the angular resolution. For the field of view, in addition to the horizontal width of the field of view, vertical height is the current technical difficulty, Nissan said that its next generation of lidar in the future vertical field of view can reach 25-40 degrees, and the current vertical field of view of everyone is generally below 10 degrees, and now the vertical field of view of Nissan and Luminar's lidar has reached 25-30 degrees. For the effective detection length, the daily production capacity reaches 300m, the current domestic hybrid solid-state lidar detection distance used by Wei Xiaoli is about 200 meters, and the lidar of Tuduitton is stronger in performance, the detection distance is about 250 meters, and the director of nissan ADAS advanced technology development department, Toru Iijima, also said that he believes that the farthest perception distance of lidar needs to be more than 300 meters to play a role on the highway. For angular resolutions currently on the market at 0.25 to 0.1 degrees, Nissan advertises future lidar angle resolutions of less than 0.05 degrees.

Joint venture learning new forces, Nissan's new self-driving program is "imitation" of Weilai?

We are familiar with the mainstream approach

Hardware layout scheme, for its software scheme, is also very clear based on the AI artificial intelligence algorithm structure, a variety of sensor redundancy perception plus high-precision map positioning of intelligent driving scheme, this set of schemes is millimeter wave radar, video head, lidar respectively processed, and then integrated for environmental integration positioning, that is, similar to the current Wei Xiaoli used scheme, different from the set of dual system double redundancy.

Joint venture learning new forces, Nissan's new self-driving program is "imitation" of Weilai?

In terms of the brain of intelligent driving - chip algorithm, Nissan did not announce this time, and it is expected to introduce Qualcomm Snapdragon Ride. The Snapdragon Ride vision system is built on a system-on-chip based on a 4nm process, integrates a dedicated high-performance Snapdragon Ride SoC and Arriver's next-generation visual perception software stack, and uses an 8-megapixel wide-angle camera developed based on a custom neural network architecture. Perception alone is not enough for cars, it also requires strong enough graphics processing power, and Qualcomm is also a leader in this field. In the new generation of Snapdragon 8 mobile platform, Qualcomm brings a new generation of mobile GPU Adreno 730, with better graphics rendering capabilities.

With the blessing of the Snapdragon Ride vision system, the vehicle's peripheral perception ability will be greatly enhanced, and the recognition of targets including lane signs, static geometries such as traffic infrastructure and guardrails, dynamic objects such as vehicles, vulnerable traffic participants such as pedestrians and cyclists, and traffic sign recognition that meets global regulatory requirements may further promote the development of autonomous driving. The system will be mass-produced in 2024, which is in line with Nissan's announcement of the development of "road reality perception" technology in the mid-2020s.

Joint venture learning new forces, Nissan's new self-driving program is "imitation" of Weilai?

This system and the DRIVE Hyperion platform with NVIDIA Orin chip as the core may become the two main solutions that car companies mainly consider when creating autonomous driving solutions in the future, and will also form a strong competitive relationship. However, the cost of Nvidia is higher, and for Nissan, which has a large number of low-end models, cost considerations must be placed in an important position, so Nissan's choice of Qualcomm, which has many years of cooperation experience, may be the best solution.

Joint venture learning new forces, Nissan's new self-driving program is "imitation" of Weilai?

From the 2019 Nissan ProPilot 2.0, we can see that it is not much different from the intelligent driving hardware and functional solutions of our new local forces in the same period. And to Nissan's latest ProPilot 3.0, we can see its hardware layout, and our new forces to build cars are almost exactly the same, but Nissan is still in the engineering car stage, from the mass production is at least 2-3 years, and our ET7 has been delivered on the road, the new forces are no longer in the experimental demonstration stage, Xiaopeng city NGP will also be officially launched.

Joint venture learning new forces, Nissan's new self-driving program is "imitation" of Weilai?

But as a traditional car company, Nissan every step of the way, the pace will certainly be more stable, but can see that its action in the joint venture brand, has been considered fast, the new forces can not be taken lightly, as a large sales of the joint venture, once it turns around, in the low, high and high levels of the car, after the full roll out of L3, L4, then the new forces with high-end main sales models, will face more challenges, then there may not be a new and old alternation, but the new and old together.

Read on