Nissan officially announced the latest development of its ProPilot advanced driving assistance system yesterday.
Before discussing the progress of this system, let’s review the history of ProPilot and look at yesterday and today’s self-driving technology of Nissan to glimpse into the future.
In July 2016, Nissan launched ProPilot 1.0, which utilized front monocular cameras and millimeter-wave radar sensors to achieve adaptive cruise control (ACC) and certain lane-keeping abilities. However, drivers needed to keep their hands on the wheel while using this feature.
Three years later, ProPilot 1.0 made its way into China, where Nissan models such as the Teana, X-Trail, and Qijun featured the system. Meanwhile, ProPilot 2.0 had been introduced in Japan as Nissan’s response to Tesla’s Autopilot.
ProPilot 2.0 features upgrades such as:
- 5 radars (1 front and 4 corner)
- 12 ultrasonic sensors
- 8 cameras (3 forward-facing, 4 AVM, 1 DMS)
- Zenrin high-precision maps
With this sensor configuration, Nissan was able to achieve high-speed navigation assistance and single-lane hands-free driving. While the former has become less unfamiliar in China, hands-free driving in a single lane is still relatively new and requires cautious optimism.
For Nissan in 2019, high-speed navigation assistance and single-lane hands-free driving were a reflection of the company’s advanced technical development. Globally, besides Tesla, no other car companies had achieved the same.
At the same time, discussions around whether to install lidar on vehicles had begun. Tetsuya Iijima, then Nissan’s General Manager of Advanced Technology Development for ADAS, said, “As a technology, lidar lacks performance that surpasses the latest technology of radar and cameras. If lidar can match the performance level of other components in the vehicle system, that would be great, but the cost and performance ratio is still not there yet.”We can infer from Tetsuya Iijima’s remarks that Nissan is not averse to using LiDAR, and if the performance of LiDAR were to improve one day, and the cost could be lowered to achieve mass production, Nissan would consider it.
The day of “lower cost and higher performance LiDAR” did not keep us waiting long, and Nissan timely launched ProPilot 3.0, with the significant change of adding a LiDAR to its latest perception architecture. The perception hardware is as follows:
- 10 cameras;
- 5 millimeter-wave radars (1 forward + 4 corner);
- 1 LiDAR (supplied by Luminar);
- High-precision maps.
It is worth noting that the position of the cameras for Nissan ProPilot 3.0 is quite special. The position of the side front cameras is similar to that of ET7, placed at the corners of the roof, forming a “trapezoid” shape. In addition, there are two cameras respectively facing forward and backward on the side fender mirrors of the test vehicle, indicating that ProPilot has redundant perception for the side front.
The supplier of Nissan ProPilot’s LiDAR is Luminar, but Nissan has stated that they are using the next-generation high-performance LiDAR, indirectly indicating that what they are currently using is not Iris. Regarding the parameters, Tetsuya Iijima, head of the Advanced Driver Assistance Systems (ADAS) development department of Nissan, said that they believe the farthest sensing distance of the LiDAR needs to be over 300 meters to be effective on highways. In addition, the horizontal FOV angle is 120 degrees, the vertical viewing angle is greater than 25 degrees, and the resolution is 0.05 degrees.
Compared with Luminar Iris, the LiDAR used by Nissan is not superior in terms of detection distance, but other data are at the mainstream level. We can also compare this LiDAR with the LiDAR used by domestic new energy vehicle start-ups.
In addition to this, Tetsuya Iijima, head of the ADAS development department of Nissan, also mentioned the topic of LiDAR installation location. Li Xiang and Xia Yiping, the two CEOs of Chinese new energy vehicle start-ups, recently commented on this issue. Tetsuya Iijima also believed that the most reasonable layout for LiDAR installation is on the top of the car, in a lookout tower-style configuration, because “the higher you stand, the farther you see,” and the LiDAR beam will not be obstructed.What functions can this L4/L5 level architecture system achieve? According to Takashi Banjima, head of Nissan’s Advanced Driver Assistance Systems (ADAS) Development Division, after adding lidar, this Ground truth perception road condition perception system can monitor road conditions in real time and quickly respond when it identifies obstacles on the predetermined route of the vehicle. We can also see the performance of test vehicles equipped with the ProPilot system in real-world scenarios in the demonstration video.
Continuous “ghost detection”
During the test, a vehicle suddenly backed up in a parking space on the side of the road and intercepted the test vehicle, but the test vehicle quickly decelerated and actively avoided into the next lane. However, there were pedestrians passing through the flower bed in the next lane, and the AEB was triggered, causing the vehicle to brake urgently.
Perhaps you might find the behavior of the vehicle in this scenario a little strange because it does not align with human driving habits where we always prioritize yielding. But don’t forget that this is a “autonomous driving” vehicle which has 360-degree perception capabilities.
- Avoiding moving obstacles on the road
In another scenario, a tire suddenly rolled towards the front of the vehicle on its planned route, which is an unsolvable problem in the present stage. However, after adding lidar, the size, shape and distance of foreign objects in the road environment can be clearly perceived, so the vehicle can make decisions based on this information. In addition, stationary obstacles on the road can also be recognized and actively avoided.
Takashi Banjima also stated that this system will still use high-precision maps. With the support of lidar, even if some parts of the map are lost, the system can continue to work, ensuring robustness.
In addition, the vehicle can actively make decisions based on road conditions in navigation-assisted driving. Nissan’s new generation road perception system seems to be a reinforced version of current navigation-assisted driving, perfectly covering “emergency situations”.
However, Nissan stated that this road condition perception system will not be completed until the mid-2020s, around 2025, and will be installed in all Nissan models by 2030. This means that consumers would have to wait at least eight years to use this system. My mood went from “anticipation” to “disheartened” in an instant.
Three brothers advance together
The entire Japanese automotive industry seems less sensitive to the global trend towards electrification and intelligence. Falling behind other regions in transformation has become a fact. However, as the saying goes: “It’s not about being bad, it’s about being worse than others.” When it comes to Nissan’s research and development achievements, we cannot help but compare the efforts of two other compatriots before we talk about their performance.
But before introducing the performance of the other two, we need to see what efforts the Japanese government has made for the development of autonomous driving.- In 2014, led by Japanese Cabinet Office, General Affairs Ministry, Ministry of Economy, Trade and Industry, Ministry of Land, Infrastructure, Transport and Tourism, National Police Agency, and major automakers such as Toyota, Honda Research Institute, Nissan, along with universities and suppliers, announced the establishment of a government-industry-academia collaborative project to promote autonomous driving on a national scale.
-
In September 2018, Japan’s Ministry of Land, Infrastructure, Transport and Tourism formulated safety technical guidelines for autonomous driving vehicles.
-
In March 2019, to accelerate the commercialization of autonomous driving and to prepare for the upcoming Tokyo Olympics, the Japanese Cabinet approved amendments to the Road Transport Vehicle Act.
-
In April 2020, Japan began implementing amendments to the Road Act that includes autonomous driving content.
Thanks to the joint efforts of government, academia, and industry, Japan’s first L3 autonomous driving car, the Honda Legend, was officially launched. Before the Legend, Audi announced that its A8 model was the world’s first L3 autonomous driving car, but due to regulatory issues, Audi’s “first” identity was not recognized. The Honda Legend, however, has been certified by the government and even has an exclusive label on the rear of the car.
But is the Honda Legend truly hands-free, and what about its configuration and price? Let’s take a look at the sensor configuration of this sensing system:
- Front dual camera;
- 5 LiDARs (2 front, 3 rear);
- 5 millimeter-wave radars;
- Relies on high-precision maps;
- Renesas R-Car V3U chip with a maximum compute power of 60 TOPS.
Yes, you read that right. This car is equipped with 5 LiDARs, using the second-generation SCALA model from Valeo, with only 16 lines, a maximum detection distance of 150 meters, and a horizontal FOV of 10 degrees. In addition, based on the installation location, it can only function as a supplementary and redundant sensor.
The most fatal flaw is that there are two prerequisites for using the Legend’s assisted driving system: high-precision map coverage is required on the usage road, and the speed limit must be below 50 km/h. The first condition limits the scope, while the second condition limits the scene. The so-called L3 is suddenly inferior and becomes a chicken rib.
However, Honda did not expect to mass-produce the Legend. This car is limited to 100 units, and users cannot purchase it outright. They can only rent it for a period of 3 years for a price of approximately 660,000 CNY. It can be seen that the appearance of the Legend is a “self-comfort” in the Japanese autonomous driving industry.
While spending money, they also do self-research
Has Honda achieved L3? Not only has it achieved it, but the government also recognizes it. However, in terms of available value, the content of Honda L3 is not high. What about Toyota?When I looked into Toyota’s autonomous driving layout, I found that Toyota has a different development path from Honda. Toyota is investing in L4 Robotaxi while increasing the popularity of L2.
Here are some of Toyota’s moves in the L4 industry (some information has been made public):
-
In 2018, Toyota established the autonomous driving R&D team TRI-AD, which focuses on autonomous driving software development.
-
In August 2018, Toyota invested $500 million in Uber. However, due to accidents, the project stalled. In 2020, Aurora acquired Uber’s autonomous driving business, and Toyota’s cooperation with Uber continued.
-
In February 2020, Toyota invested $400 million in the domestic autonomous driving start-up Pony.ai.
-
In July 2020, Toyota’s autonomous driving team TRI-AD announced the establishment of an independent company, Woven Planet. Recently, Woven Alpha, a subsidiary of Woven Planet, has started to develop a low-cost autonomous driving solution that follows Tesla’s pure vision approach on the direction of autonomous driving development.
-
In 2021, Woven Planet participated in the C round financing of autonomous driving company Nuro, and the specific amount was not disclosed.
-
In April 2021, Toyota acquired the Lyft autonomous driving department for $550 million.
Toyota has chosen a strategy of self-research and investment and acquisition in the development of autonomous driving, spreading the risks and increasing the success rate.
However, L4 is ultimately high-end, and on the one hand, Toyota is also working hard to increase the popularity of L2. If you look at the configuration table of Toyota models, even low-end models are equipped with the TSS 2.0 system, which can provide lane keeping, adaptive cruising, and other functions for drivers. Although this system cannot compare with the ADAS system developed by new forces in terms of experience, it is commendable for Toyota to increase the popularity of L2 auxiliary driving and active safety configuration through actions alone.
As for more advanced assisted driving, Toyota is not unable to do so, but is currently limited by factors such as cost, and general users cannot experience it.
Toyota provides a set of assisted driving system called Teammate for overseas markets’ Lexus LS 500h and Mirai top users, which has turn signal lane-changing and navigation-assisted driving functions, and is also an L2 system in terms of functionality and capability.
By comparison, the Japanese trio is making progress in autonomous driving, but Nissan’s 2030 vehicle still makes people gasp for breath.Domestic new forces don’t like to evaluate their capabilities using the SEA standard, but landing scenes are the most realistic reflection of their capabilities. This year, the top three new forces have focused on landing city navigation assisted driving, reflecting stronger assisted driving capabilities with lower takeover rates. The next one is to connect high-level assisted driving from point to point.
Judging from the intensity and pace of the new forces’ “internal competition,” it seems that Nissan’s 2025 R&D landing plan and 2030 on-road vehicle production plan have already announced failure in China, even though these time points refer to the R&D and production pace in Japan.
In conclusion, I always believe that assisted driving should not pursue “dazzling” under special conditions, but should strive for “silent perfection” in daily driving. Regarding the road of automatic driving, Honda took the lead in mass producing L3, but this system has gradually deviated from the essence of “assisting users in driving.” Toyota is based on L2 and has been continuously attacking L4 on the development path without any flaws, but it has been unable to provide users with low-cost, good experience assisted driving for a long time.
With a brand-new architecture, Nissan has achieved navigation assisted driving while endowing this system with stronger AEB capabilities, but it is still in the testing stage. From the perspective of its 2025 and 2030 strategic plans, if Nissan wants to gain a foothold in China’s main battlefield of assisted driving, it still needs to make every effort to advance.
This article is a translation by ChatGPT of a Chinese report from 42HOW. If you have any questions about it, please email bd@42how.com.