Text | Angel Lee
Edit | Su Jianxun
Another startup is working on both car-making and autonomous driving.
Recently, Xingjiao Technology, co-founded by Zhao Ruixuan, former vice president of Xiaoma Zhixing, and Wang Qingzhou, former technical head of Waymo's architecture department, announced the launch of an L4 level autonomous driving pure electric van heavy-duty truck logistics vehicle Apebot I more than 7 months after its establishment.
Founded in August 2021, Xingjiao Technology intends to integrate the wire-controlled electric chassis with automatic driving technology, serve the logistics industry with unmanned driving technology, and create an intelligent electric transportation network. The team has secured millions of dollars in strategic financing.
The two co-founders are from self-driving companies Xiaoma Zhixing and Waymo.
CEO Zhao Ruixuan was previously the vice president of Xiaoma Zhixing, a self-driving company, and served as an executive in Facebook, Sina Weibo, Lenovo Group and other companies. During his time at Pony Chi Heng, he helped Xiao Ma Zhi Xing complete the cooperation and financing of a number of car manufacturers including Toyota and Tier 1.
CTO Wang Qingzhou is the former technical leader of Waymo's architecture department and has served as a long-term technical leader at Google's Machine Learning Research Institute. Graduated from the Department of Automation of Tsinghua University, he entered Google's US headquarters in 2011, joined Waymo Architecture in 2017, and was responsible for the research and development of multiple core technologies for unmanned vehicles during his tenure, in addition to working at Uber headquarters.
At present, the company has technology research and development centers in Silicon Valley, Shanghai and Guangzhou, China, and the team has talents in chassis, electric drive, battery, wire control, chip design, etc., and the members are from Google, Xiaoma Zhixing, Qualcomm, Xiaopeng Automobile, Bosch, ZF and other enterprises.
For example, Ma Junye, vice president of vehicle engineering, has nearly 15 years of experience in the vehicle research and development industry, and has served as the chassis director of Xiaopeng Automobile and the director of the chassis department of BAIC; the person in charge of perception, Zhao Qijie, was previously the core person in charge of the mass production perception technology of Xiaoma Zhixing.
At present, self-driving trucks can obtain rapid commercial landing, which has almost become the consensus of the industry.
Not only Waymo, Baidu, Xiaoma Zhixing, etc. are doing related explorations, tucson, which has been listed, has more than 20 customers and nearly 7,000 self-driving truck reservations in the United States, which also confirms the feasibility of the road to a certain extent.
Zhao Ruixuan also told 36Kr: Compared with passenger cars, trucks as 100% operating tools, their unmanned urgency is higher than passenger cars, and the use of trucks is more at high speed and has fixed routes, which also makes automatic driving land faster.
She described the ideal entrepreneurial endgame to 36Kr: "Our ultimate model should be to become an unmanned electric transportation provider. Unmanned trucks shuttle on the highway to do intelligent distribution. Based on electric chassis technology, our back-end management system can monitor the status of all vehicle parts and autonomous driving, better dispatch freight and charge vehicle maintenance. ”
Therefore, in addition to building an autonomous driving system and exploring an intelligent logistics system, Xingjiao Technology also intends to lead the definition of the next generation of trucks.
"The end game of the truck must be realized by electrification + automatic driving. If you only do electrification, the economic advantage of the truck is not so obvious, because the driver cannot be removed. Only through automation + automatic driving, removing the cost of the driver, and removing the cab in the vehicle hardware, can the economy of the logistics truck be maximized. ”
Source: Xingjiao Technology
On the one hand, it is the economic needs of logistics trucks, and on the other hand, it is also the need for the development of autonomous driving. In the process of interviewing 36Kr, Zhao Ruixuan stressed many times: only by owning the IP of the vehicle can you control all the data of the car.
"If there is no car, just make a software system, and then do simple integration with the car company, in fact, you can't get all the data in the car, which also leads to the need for autonomous driving companies to continue to deploy fleets and collect a large number of real road test data."
CTO Wang Qingzhou told 36Kr: In the past, the industry relied on fleet-scale road testing, which did make automatic driving develop to a certain extent. But it is not feasible to solve the corner case (long-tail scenario) by the heap engineer writing the rule. Vehicles may run on the road for a long time to encounter good scene data, and the cost of this process is uncontrollable.
Therefore, Xingjiao Technology believes that a large amount of data can be obtained by building cars, selling cars, and operating fleets. The data-driven approach enables iterative autonomous driving systems to eventually become unmanned.
It can be understood that what Xingjiao Technology emulates in the field of trucks is exactly the route of Tesla in the field of passenger cars: first electrification, then autonomous driving.
Building a car is the only way to go
However, this does not mean that Xingang Technology should make the construction of trucks particularly heavy.
It is understood that the model adopted by Xinglan Technology is to customize vehicles with car companies, similar to the joint development of D1 models by Didi and BYD. Zhao Ruixuan told 36Kr: In addition to the automatic driving system, the team will also focus on the research and development of the whole vehicle, the three-electric and wire control, the definition of the vehicle electronic architecture, and the rest of the chassis, batteries, etc. are achieved through cooperation with the main engine factory.
"Half of the team is people in the autonomous driving industry, and half of them are people related to the car and the industrial chain. It is expected to reach 50 people during the year. Zhao Ruixuan said.
On the occasion of its establishment more than 7 months ago, Xingjiao Technology released its pre-installed L4 level automatic driving pure electric van heavy-duty truck Logistics Vehicle Apebot I, as well as the automatic driving software system APANet.
The overall 4x2 van, the total weight of the car and cargo is 18 tons, the length of the container is 9.6 meters, and the maximum volume of the container can be 67 cubic meters, which is currently commonly used in the logistics industry in the province. The Xinglan Technology program focuses on the short- and medium-distance logistics and transportation market, including the last mile.
Zhao Ruixuan told 36Kr that Apebot I was jointly built by Xingjiao Technology and a central enterprise background truck OEMs, and was equipped with a wire-controlled chassis solution developed in cooperation with International Tier 1.
On the basis of the electric wire control chassis, the software level can also integrate the vehicle domain controller and the battery BMS system, and the status of the core components such as battery and electric drive is uploaded to the cloud platform, becoming an important part of the logistics fleet management system.
According to Wang Qingzhou, in terms of automatic driving, Apes I has 9 cameras, 3 millimeter-wave radars, and 3 solid-state lidars, with a total of 15 sensors, covering an area of 250 meters forward and 150 meters backward.
Ma Junye, vice president of vehicle engineering of Xingjiao Technology, said that Apebot I's NEDC cruising range can reach 450Km, and the actual full load endurance reaches 380km, which can meet most of the cross-city transportation needs. The model is currently in preparation for various validation tests prior to mass production, and production is expected to begin in February 2023.
"There are now small batch orders in talks, and the prototypes are basically ready and mass-produced at the end of the year." After small batch production next year, we will leave part of the capacity in hand and sell part to the logistics side. Zhao Ruixuan said.
Xingjiao Technology said that the sensors of these vehicles will be pre-installed, and the vehicles sold to the logistics party will release some L2+ level assisted driving functions in the early stage. The internal self-built operating fleet will be tested in L4 autonomous driving state.
"When we feel that some sections of the road are more reasonable, we can upgrade the OTA and open up and achieve unmanned driving." Wang Qingzhou said.
In addition to the first car, Xingjiao Technology also gave a strategic plan for the development of several models in the future.
"The first generation of cars mainly changed from fuel vehicles to electric vehicles, covering short- and medium-distance logistics and transportation scenarios. In the second generation of cars, we will do more content for autonomous driving, realize the removal of drivers, and build an intelligent electric transportation network. ”
It is expected that in 2025, the third generation of xingang technology will remove the cab of the truck, because the current truck cab cost can account for 30% or 40% of the vehicle, coupled with the removal of human driving, the production and operation of the truck will have a huge optimization space.
Previously, the National Highway Traffic Safety Administration (NHTSA) said it was proposing changes to U.S. traffic regulations, and that autonomous vehicles without steering wheels and accelerator pedals would be allowed to be tested on the road in the future. "This is a very clear signal and trend." Zhao Ruixuan said.
Although it is not too early to enter the game, Xingjiao Technology said that this entrepreneurial time node is not too late, because the team can stand on the shoulders of the industry and adopt the latest technological achievements: such as the use of deep learning algorithms, strengthen the 3D detection of objects by visual technology, and weaken the dependence on high-precision maps; in terms of hardware, the latest components of the industry such as solid-state lidar can be used. "The historical baggage of autonomous driving systems will not be so heavy."
Of course, self-driving truck companies want to lead the car manufacturing, which is a road that no one has traveled so far, and it also needs to invest more than its peers. As a latecomer, Xing orangutan Technology still has a long way to go.
The following is an excerpt from the interview with Zhao Ruixuan and Wang Qingzhou from 36Kr:
36Kr: Why did you choose to start a business at this juncture? Why are trucks built?
Xingjiao Technology: Technically speaking, software and hardware (wire control, self-driving calculation system, sensor) have almost matured in the past few years, and many can reach mass production in 6 to 12 months. Battery technology has also been accumulated and improved over the years, making truck application scenarios a reality. Now is the golden time to do trucks.
36Kr: At present, everyone's demo in the field of self-driving trucks is doing well, what is the core competitiveness of the walking ape? What are your focus on technology and products?
Xingang Technology: Many companies today's demos are actually waymo that many years ago, but large-scale landing is still not achieved. So business models and data-centric R&D models are critical.
We have developed pure wire control chassis technology since day 1, and the combination of soft and hardware is electrified + intelligent. The commercialization of autonomous driving needs to rely on hardware and operations to land. Robotaxi would not have been able to operate at scale without front-loading integration of vehicles. The same goes for trucks, where to achieve large-scale robotruck, you have to touch the vehicle's hardware. But this does not mean that every link of car manufacturing must be done end-to-end, but it is necessary to have the understanding of vehicle research and development, the ability to define the core three electricity and wire control, and E/E architecture.
36Kr: What do you think of the current problem that the commercialization of self-driving trucks has not really run through?
Xingjiao Technology: I think the biggest problem is that it skips the electrification of trucks and goes straight to intelligence. This is very different from a passenger car.
Emerging self-driving technologies must also be built on the basis of electrification. The next-generation cars of Waymo, Cruise, Nuro and others are all electric. The solution based on the traditional fuel vehicle + transmission has a very large reaction delay to the system and is linearly poor. In contrast, the control technology of the motor is very mature, the response is fast, the linearity is good, and it is easier to achieve a stable and safe control system. So it's going to be electrified first, and then it's going to be autonomous.
36Kr: There are not many cases of startups leading car manufacturing, will OEMs be willing?
Xingjiao Technology: Trucks are actually much lighter than passenger cars, and will not be as expensive as passenger cars. There are several reasons: trucks are not a consumer product like passenger cars, and a lot of investment is invested in exterior interior and chassis design. The truck is essentially a cab, the chassis and the cargo box are assembled together, and it is a production tool, the demand for modeling is not large, the production line will not be as complex as the passenger car, and the investment will be much smaller.
Our team has a strong accumulation of industrial cooperation, not only understands the automatic driving software and hardware, but also understands the car, and can use the resources of existing car companies and supply chains to define the vehicle. We are not to subvert the original car factory, China's truck production capacity is very sufficient, the domestic supply chain is also very complete, we will be in the original supply chain, to the greatest extent possible to cooperate and win-win