Car stuff (public number: chedongxi)
The author | James
Edit the | Xiao Han
Toyota seems to have become the most anxious company in the era of autonomous driving.
Nowadays, autonomous driving is almost an iteration every day, and the Toyota Group, which has the second highest market capitalization and the first sales volume in the world, is naturally indispensable to the layout in the field of automatic driving.
In recent years, Toyota's public investment and mergers in the field of autonomous driving have exceeded 1.68 billion US dollars (about 10.7 billion yuan), launched five generations of unmanned vehicles, and countless cooperation and research and development investments... Among the traditional car companies, it is almost impossible to find a second company like Toyota to devote money to autonomous driving.
However, Toyota has relatively few automatic driving results that it can really do. For example, there are not many models in Toyota and its brands equipped with L2 autonomous driving systems, but only two lexus models are equipped with L2 automatic driving advanced functions such as command lane changes. Not to mention the noa such features, in Japan and the United States only Toyota Mirai Limited, Lexus LS500h two top-of-the-line models of the top configuration can be selected.
▲Toyota L2 automatic driving assembly summary
Obviously, after a large amount of investment in research and development, Toyota has few obvious results in the field of autonomous driving.
In terms of technology routes, Michael Benisch, vice president of engineering at Toyota's Woven Planet, was recently interviewed, and the meaning of the words was to develop a purely visual technology route. This is in sharp contrast to Toyota's 10-year vision + radar technology route, and highlights Toyota's anxiety in the field of autonomous driving.
So, what are Toyota's latest autopilot achievements, why do you say that Toyota has a strong anxiety about automatic driving, and how anxious is Toyota? Through in-depth investigation and research, Che Dongshi found the answer to the question.
First, smash the layout of tens of billions of Chinese and American companies have invested over and over again
As early as around 2016, Toyota has begun to invest in autonomous driving around the world, and it is almost always a startup. In 2018, Toyota began a large-scale investment and research and development of autonomous driving.
In March of that year, Toyota officially established the toyota research and development team TRI-AD (Toyota Research Institute-Advanced Development, Inc.), with the goal of providing software for autonomous driving and ultimately realizing the collaborative work between people and vehicles to create safe, convenient and efficient travel services.
In August 2018, Toyota invested $500 million (about 3.2 billion yuan) in Uber and expanded cooperation in the field of autonomous driving. In the process of cooperation, Uber uses Toyota Xena MPV to build L4 autonomous vehicles and integrateSure and Toyota's respective autonomous driving algorithms to control vehicles.
▲ Toyota's cooperation model with Uber
However, due to a serious traffic accident that killed passers-by during Uber's self-driving test, its self-driving test had to be suspended from 2018. In 2020, before Uber sold its self-driving business, it failed to build a self-driving car based on the Xena MPV.
Autopilot company Aurora "revived" its partnership with Toyota after acquiring Uber's self-driving business, finally launching a saena-based driverless car in September 2021. In March 2022, Aurora plans to start operations by the end of 2024 based on the unmanned vehicle test built by Toyota Xena.
▲Aurora is a driverless car based on Toyota Xena
It's been three and a half years since Toyota invested in Uber's autonomous driving.
After seeing that Uber's self-driving tests were almost impossible to carry out because of the accident, Toyota had already set its sights on other self-driving companies.
In February 2020, Toyota invested $400 million (about 2.5 billion yuan) in domestic autonomous driving startup Xiaoma Zhixing, which was also one of the largest financings in the domestic autonomous driving field that year.
In fact, in the domestic automatic driving test, Xiaoma Zhixing has adopted a considerable number of Lexus RX450h models, and the sixth-generation automatic driving system released in January this year is based on Toyota Xena.
▲Toyota Syna Autono-MaaS equipped with the sixth generation of the Pony Chi Heng autonomous driving system
It can be said that there is a fairly close connection between Pony Zhixing and Toyota.
By the second half of 2020, the competition for autonomous driving is becoming more intense. Waymo removed the human security officer and let the driverless car run on its own. Since 2021, the commercial operation and unmanned testing of domestic autonomous driving are gradually landing.
Toyota has not stopped, and in July 2020, Toyota's TRI-AD announced that it has entered the next stage of development after two years of development and established a holding company, Woven Planet. The holding company has three subsidiaries, Woven Core, Woven Alpha and the Woven Institute.
Among them, Woven Core specializes in the development of autonomous driving technology and explores commercialization; Woven Alpha is exploring new projects to incubate Toyota, including the smart prototype city Woven City, autonomous driving software Arene, automatic mapping platform AMP, etc.; Woven Institute is an investment fund that specializes in investing in autonomous driving, AI, robotics and other industries in the industry.
▲Woven Planet Holdings Company Structure
This is a big step for Toyota in the field of autonomous driving.
In March 2021, two months after the new company was officially registered, Woven Planet invested in U.S. self-driving company Nuro, the first since the new company's inception. As part of Nuro's Series C financing, Toyota's specific investment has not been announced.
Then in April 2021, Toyota acquired the autonomous driving division of another U.S. ride-sharing giant, Lyft, for a $550 million deal. Among them, $200 million (about 1.3 billion yuan) is an advance payment, and $350 million (about 2.2 billion yuan) toyota will pay Lyft in five years.
In 2021, the newly formed Woven Planet will have new developments in the field of autonomous driving almost every month. In addition to Nuro and Lyft's autonomous driving divisions, it has invested in Ridecell, a self-driving mobility platform company, CARMERA, a high-precision mapping company, Renovo Motors, an automotive operating system company, and UP.Partners, a venture capital fund that has invested in self-driving startups for many years...
It can be said that before 2021, Toyota cautiously invested in several representative autonomous driving companies in the industry, and since 2021, it has invested in and acquired a bunch of autonomous driving companies on a large scale.
Faced with the rapid iteration of the autonomous driving industry, Toyota seems to have panicked.
Second, L2 is basically popular, but it still can't enter the first echelon
Toyota Group's brands Toyota, Lexus, and Subaru have passenger car layouts in China, and the car thing statistics show the popularity of L2 autonomous driving of Toyota's models, and find that it is not low. However, advanced L2 autonomous driving functions such as NOA and lane change assist have a fairly low assembly rate.
Overall, as long as it is based on Toyota's TNGA platform to build a model, do not buy the minimum configuration basically have L2 level automatic driving, no L2 automatic driving model for sports cars and models below 100,000 yuan.
Calculations found that among the 33 car series of the Toyota Group, the assembly rate of L2 level autonomous driving reached 74.5%.
▲Toyota L2 automatic driving assembly summary
However, if you look closely, you will find that the models equipped with L2 automatic driving functions are basically Toyota TSS2.0 (Toyota Safety Sense 2.0) system, which only has a single-lane L2 automatic driving function.
If you want to buy a model with a command lane change function, only the lexus NX and LS models can be used with high equipment, it can be said that the command lane change has become a niche function.
In overseas markets, Toyota offers an L2-level autonomous driving system called Teammate, which has NOA functions, but only the Toyota Mirai top version Limited and Lexus LS500h are eligible for optional.
▲ Lexus LS500h equipped with Teammate
In fact, the hardware configuration of this Teammate automatic driving system is very good, with a total of 4 lidars in front and rear and left and right fenders, 2 single sight cameras, 5 millimeter wave radar, and 2 rearview cameras for L2 automatic driving. There are also 4 surround view cameras and 12 ultrasonic radars for automatic parking.
▲Toyota Teammate sensor configuration
The system, combined with redundant computing chips, is available for $5,170 (about 33,000 yuan) and has a 10-year subscription to autonomous driving software.
Functionally, the system enables L2-level navigation assisted driving on high-precision maps-enabled highways, automatic lane changes, and allows drivers to drive off their hands when driving in a single lane.
▲It is equipped with teammate Toyota Mirai
It can be said that this system is very strong in both price and function, supports off-the-hand driving of few models, and the optional price of perception computing equipment is not particularly expensive.
After the optional installation was completed, the price of Toyota Mirai Limited reached $71,000 (about 455,000 yuan), and the Lexus LS500h came standard with Teammate, which reached a price of $113,000 (about 722,000 yuan).
According to the previous analysis, Toyota modified and tested based on the Lexus LS when developing the Teammate automatic driving system, so it is not surprising that the first models equipped with the Teammate automatic driving system are the Lexus LS and the Toyota Mirai on the same platform.
However, if you want to expand to more models, I am afraid it is still a long process. Even before the mass adoption of Teammate, Toyota struggled to achieve the massive amounts of data necessary for pure vision (such as visual images on the sides and rear), making it more difficult to iterate on autonomous driving.
Perhaps, how to iterate is the core of Toyota's self-driving anxiety.
Third, L4 is mainly out of the car, and its own fleet has no large-scale application
In fact, Toyota began to develop autonomous driving earlier, launching unmanned vehicles in 2013, which has been nearly 10 years. But in this nearly 10 years of development, you may have seen unmanned cars with Toyota/Lexus logos, but these cars are all vehicles provided by Toyota to self-driving companies, and Toyota's own unmanned vehicles are rarely on the road.
1, research and development of nearly 10 years finally have NOA
At cess in January 2013, Toyota showed off a self-driving car based on the Lexus LS600hL, which was equipped with a 360-degree lidar on the roof, as well as three cameras and millimeter-wave radar to perceive the surrounding environment, which was also the beginning of Toyota's self-driving development.
▲Toyota's first generation of unmanned vehicles
After four years of research and development, Toyota launched its second-generation autonomous driving test vehicle, the Platform 2.0, in March 2017. And Toyota has clarified the research and development route for autonomous driving in this generation of models, in addition to developing L4 (Chauffeur), but also developing L2 (Guardian) based on the same platform.
Toyota Platform 2.0
Soon, Toyota launched the Platform 2.1 platform in September 2017, with basically unchanged sensor layouts, but the interior of the car changed from one steering wheel to two. Toyota noted that the two steering wheels are capable of testing the way human drivers take over autonomous vehicles.
Toyota Platform 2.0
At CES in January 2018, Toyota launched the iterative Platform 3.0 autonomous driving test vehicle, and the sensor layout of this car is rushing to mass production.
It can be seen that the Platform 3.0 has a short-range lidar on the front, rear, and left and right wing panels, and has 4 long-range lidar and 8 millimeter-wave radar on the roof, which has a relatively high degree of similarity with the current 4-lidar mass-produced products.
▲Toyota Platform 3.0 sensor layout
A year later, at CES 2019, Toyota unveiled the Platform 4.0. There are two major upgrades to the model, one is the upgrade from the fourth-generation Lexus LS to the fifth-generation model, which is built on Toyota's TNGA-L platform; the second is that the autonomous driving system is powered by the hybrid system's power battery instead of a 12V battery.
Toyota Platform 4.0
These two upgrades bring Platform 4.0 closer to mass production.
Sure enough, in 2021, Toyota launched the Lexus LS500h and Toyota Mirai (on the same platform as the two Lexus LS) in Japan, and the Teammate autonomous driving system was mass-produced, both in terms of sensor layout and electrical and electronic architecture, which have relatively high similarities with Platform 4.0.
2, L4 passenger car has not yet landed
However, in the field of L4 autonomous driving, Toyota has achieved even less.
At the 2019 Tokyo Motor Show, Toyota launched a pure electric minibus called e-Palette, which has L4 level automatic driving capabilities. At the 2021 Tokyo Olympics, 20 e-Palettes will be used as shuttle buses in the Olympic Village.
However, even running in a closed park like the Olympic Village, e-Palette had an accident and hit a visually impaired athlete, resulting in the athlete's regretful withdrawal.
▲Toyota e-Palette concept car
After the accident, Toyota stopped operating the e-Palette in the Olympic Village.
This year, Toyota also brought the e-Palette to China. According to domestic media reports, from January to March this year, Toyota e-Palette conducted manned experiments in Shougang Park. During this period, the e-Palette operated for a total of 15 days, carrying more than 13,000 passengers.
However, in addition to this e-Palette, Toyota does not have a public-facing L4 passenger model.
At the same time, according to the data released by the California Vehicle Management Bureau (DMV), Toyota's autonomous driving test test in California is not active, and the total test mileage of the whole year ranks third among the vehicle companies, ranking after the general Cruise and Mercedes-Benz; if you add other autonomous driving companies, Toyota's total test mileage ranks 13th, and only 4 cars run on the road.
And the MPD value that tests the ability to drive itself, Toyota ranks even lower, reaching 20, and every 33.32 miles it needs a driver to take over, which is really not a good result. Moreover, compared with the previous Mercedes-Benz, the gap between Toyota is still very large.
After nearly 10 years of research and development, Toyota has produced mediocre results, L2 level NOA has just been delivered in the Japanese and American markets, and L4 autonomous driving field has few achievements.
These results are also the result of Toyota's huge investment and long-term research and development. This has to make people think that Toyota's automatic driving has indeed encountered a considerable bottleneck.
Fourth, the development path is not clear where Toyota is driving autonomously
Looking at Toyota's autonomous driving development path in the past 10 years, it can be found that when L2 and L4 began to land on a large scale, Toyota's autonomous driving development path was still unclear.
From the perspective of research and development and investment, Toyota has not found the focus of development.
On the one hand, Toyota has developed four generations of self-driving test vehicles and wants to develop its own autonomous driving technology; on the other hand, it has invested in a large number of autonomous driving companies. Moreover, Toyota has achieved not many results under the premise of both research and development and investment.
From the analysis above, it can be found that Toyota' self-developed L2 or L4 achievements are not the head of the industry, or the first to land, so the results are still quite far from the head companies of automatic driving technology.
In terms of investment, Toyota has invested more than $1.68 billion (about 10.7 billion yuan) in autonomous driving. However, Toyota's relationship with various autonomous driving companies is not clear, basically to provide vehicles, and more in-depth operation and mass production cooperation has not yet been announced.
In contrast, General Motors, which is also the world's leading car company, and Honda, which is also a Japanese car company, the autonomous driving routes of these two companies are relatively clear.
In particular, General Motors, on the one hand, took the lead in the industry to mass-produce L2-level autonomous driving that can be taken off, on the other hand, its self-driving company Cruise has achieved manned operation and unmanned testing in the United States, which is on the same level as Waymo, the "boss" in the field of automatic driving.
▲Cruise Origin unmanned vehicle
In the foreseeable future, GM can promote the large-scale commercialization of cruise unmanned vehicles, while delegating L4 technology to mass production passenger cars.
Honda will become the world's first mass-produced L3 level autonomous driving car company in 2021, although the production is limited and it has not gone out of Japan, but the significance of "the first mass production L3" is quite significant. Obviously, Through autonomous driving self-development, Honda will gradually transition from L2 to L3 and then to L4, which also has a clear route.
▲The Honda Legend is the world's first mass-produced L3 autonomous vehicle
In addition to the lack of focus and clear route at the level of R&D investment, Toyota's technical route for autonomous driving is also facing changes.
Not long ago, Michael Benisch, vice president of engineering at Toyota's Woven Planet, said in an interview that the use of components such as lidar during testing is very expensive and difficult to expand. At the same time, he also said that Toyota's self-driving research and development needs a lot of data.
This passage is considered to be Toyota's bet on pure visual self-driving.
Obviously, the description of Michael Benisch, vice president of engineering at Woven Planet, overlaps with Tesla's pure visual route.
▲Tesla adopts a pure visual autopilot solution
However, pure vision is not the same as Toyota's nearly 10 years of lidar + camera sensor combination. If you start to change the sensor solution from now on, Toyota may waste a few years of time.
A few years from now, the self-driving industry will surely see a different story.
In fact, there is no right or wrong in the technical route of autonomous driving, but under the existing technical route, the whole system is made more reliable. If after many years of research and development on a technical route, it will be quite time-consuming and laborious to tear down and restart the new technical route.
From a mass production perspective, toyota's mass production autonomous driving system does not have a clear timeline. For example, leading traditional car companies such as General Motors and Volkswagen Group have announced the mass production schedules for L2 and L4, and promoted the mass production time point of various technologies according to the mass production schedule.
This is relatively more transparent for both the industry and the average consumer. However, Toyota has rarely announced the progress of research and development and mass production of autonomous driving technology, which may also be another portrayal of Toyota's anxiety in the field of automatic driving.
Conclusion: Toyota autonomous driving still has a lot of anxiety
As traditional car companies continue to accelerate the pace of electrification and intelligent transformation, the future track competition will be more intense, and more high-quality products will be launched in turn.
In the major car companies hardware foundation and software tuning level is getting higher and higher today, intelligent driving system is reducing the burden on drivers, improve driving safety. But the competition has also become more intense, and more L2 self-driving functions, or products that can be differentiated from other car companies, can win the love of consumers.
Obviously, Toyota's autonomous driving still has a long way to go, and anxiety in the development of autonomous driving will always accompany Toyota.