In the first three articles of this series, "Autonomous Driving Middleware Part 1: Is AUTOSAR being "marginalized"?" , "Autonomous Driving Middleware Part 2: Communication Middleware, DDS and SOME/IP, who will dominate?" , and "Autonomous Driving Middleware Part 3: Inventory of Active Players" , we have briefly introduced dozens of active players in autonomous driving middleware, but there is still a very important force that has not been mentioned, and that is the OEM.
OEMs such as Volkswagen, Toyota, Daimler, Ideal, and Xiaopeng are all developing their own middleware. In fact, some OEMs even refer to their own middleware as "self-developed operating systems".
Of course, there is controversy over whether it is necessary for OEMs to develop their own middleware. In addition, among OEMs that choose to develop their own middleware, the degree of their own development varies.
1. Supplier: OEMs do not need to develop their own
Suppliers generally believe that OEMs only need to purchase middleware from them and there is no need to develop their own. The main reason is that in the eyes of these suppliers, middleware is a relatively common product and is not the focus of OEMs to achieve differentiated competition.
Bi Xiaopeng, co-founder of Huayutongsoft, said: "Middleware and other low-level technologies can help OEMs reduce the difficulty of developing upper-level software and improve development efficiency, but end users do not care about what technology is used in the bottom layer of autonomous driving. They are more concerned about the application layer."
Therefore, Bi Xiaopeng believes that OEMs should also focus more on matters that can demonstrate their competitiveness to consumers - such as upper-level application software, which are tangible and visible to ordinary consumers.
Cao Bin, general manager of Neusoft Reach, also said in a conversation with Zhao Fuquan some time ago: "Middleware will become more and more mature, and will eventually form a set of widely used standardized software, combined with corresponding management tools and adaptation services. Middleware is a common component and should be widely used in many models of many OEMs. The more widely used, the better."
Cao Bin believes that middleware should be developed by independent suppliers. “This is because independent suppliers have stronger professional capabilities and it is easier for independent suppliers to cooperate with multiple OEMs, thereby expanding the scope and scenarios of software applications.”
Cao Bin admitted that some OEMs would like to develop the underlying universal basic components themselves in order to achieve more thorough differentiation. However, he believes that this is not necessary, "because the input-output ratio of this strategy is too low. It is equivalent to re-developing most of the universal software in order to achieve some minor differences in the product."
However, during that conversation, Zhao Fuquan asked Cao Bin a question: If Neusoft Rich did the same thing for many OEMs, would it eventually lead to the homogenization of these companies' products?
Cao Bin's answer to this is: For most models, these basic functions do not need to be differentiated, but must be universal; at the same time, they must be able to maintain long-term stability and reliability, so that they can play their due role. Otherwise, the vehicle will have problems.
Behind this controversy, there is actually a question: Can a basic technology like middleware become a competitive barrier for an OEM?
In October 2021, Horizon Robotics founder and CEO Yu Kai said in a conversation with Zhao Fuquan:
“According to my personal observation, no company has ever been able to form a long-term competitive barrier or moat by relying on a new technology, because valuable new technologies will eventually be industrialized and popularized no matter how bumpy the application process is. Moreover, the best user of a technology is often not its inventor. For example, the tank was invented by the British, but the Germans used it better and it became the core of the ‘blitzkrieg’.
"Of course, technology inventors have first-mover advantages, and companies hope to seize the market by virtue of their first-mover advantages. However, whether this first-mover advantage can be brought into play depends on whether the company's purpose is clear. In fact, if a company has a clear understanding of the relationship between products, brands and users, even if it is not the inventor of the technology, it can still find technology resources to apply; if it has not figured it out, even if it has a first-mover advantage in technology, it will not be able to realize it."
Regarding basic, platform, and common technologies, Yu Kai’s views are similar to Cao Bin’s: OEMs do not need to do it themselves, and the best solution is to leverage external ecosystems to achieve a “leverage” effect.
2. Most OEMs: It is necessary to develop their own
In addition to suppliers, there are also people in the OEM camp who believe that there is no need for OEMs to develop their own middleware.
For example, the head of a domain controller of a second-tier new car manufacturer said: "A rule in the software development industry is to avoid reinventing the wheel, which also applies to automotive software development. For OEMs to master the ability to develop autonomous driving themselves, the core is not how to implement the details of the middleware. If there is a mature, reliable, stable, and highly cohesive commercial middleware module, external acquisition is a better choice."
However, most OEMs do not agree with the view that "middleware is a universal product and therefore does not need to be developed independently."
The director of perception algorithms at a new car manufacturer said: "The function and quality of the middleware have a great impact on development. For example, whether the video in the perception link can be perfectly replayed depends on whether the middleware supports the playback function. For example, if the middleware is incomplete or occupies too much memory resources, it will also bring a lot of extra burden to development."
The head of the autonomous driving domain controller of another new car manufacturer said: "There is currently no unified standard for middleware. Each company customizes it. Why do you say 'no need to develop it yourself'?"
The person in charge of autonomous driving at a joint venture OEM said: "I think middleware is an important means to realize SOA (service-oriented architecture), and SOA needs to have different functions for different people. To meet this demand of SOA, middleware naturally needs to be differentiated. Therefore, self-development is necessary."
The chief scientist of an autonomous driving subsidiary of a Chinese automaker said: "Middleware's data communication, resource management and task scheduling have a huge impact on autonomous driving. The scenarios faced by each company's autonomous driving system and the speeds they need to achieve are very different, so the demand for middleware will also vary a lot."
Although it is true that the middleware itself has the function of "software and hardware decoupling" and is therefore not "strongly bound" to the application scenario, the complexity of different scenarios varies, and the requirements for middleware will still vary greatly.
Specifically, even on the same car, the three scenarios of highway, urban road and autonomous parking have different data scales and vehicle speeds, so the data throughput and latency requirements of the middleware are also very different. Taking the communication middleware DDS as an example, the QoS (service strategy) required by the above three different scenarios is obviously different.
Previously, Volkswagen Wenwen CEO Zhang Renjie, who was once the general manager of QNX Greater China, said in an interview with the author: "The difference and advancement of technology of one OEM compared to another OEM depends to a large extent on your overall framework, and the framework depends on the combination of OS middleware. Various different middlewares ultimately constitute your basic framework."
In addition to the OEMs, there are also some upstream suppliers who believe that middleware is important for improving the competitiveness of OEMs and therefore deserves attention.
For example, the marketing director of a domain controller manufacturer said: "Well-made middleware can help complete the adaptation of various lower-level hardware. If the lower-level hardware needs to be switched in some cases, the workload of modifying the upper-level application will be very small. In the long run, for OEMs, good middleware means more controllable costs."
According to the opinions of the above-mentioned people, for OEMs or autonomous driving companies, self-developed middleware is not an option but a must.
3. Even if it is really necessary, it is difficult for OEMs to develop their own
However, due to objective difficulties such as personnel training, team integration, and R&D cycle, developing self-developed middleware is not an easy task, especially for traditional OEMs.
Challenges in team integration and organizational culture are easy to underestimate, but traditional OEMs have come from the previous era, and conflicts between software and hardware and between new and old things are inevitable. Such conflicts can easily make software talents feel hesitant.
The CTO of an autonomous driving company once told me: "Some time ago, an autonomous driving subsidiary of a certain OEM invited me to join. I asked a question: 'What is the relationship between you and XX Research Institute? XX Research Institute is engaged in autonomous driving. Is it because they failed that you started a new one? Why did they fail? Is it a system problem? Will your system be better?' The other party didn't know how to answer me, so I didn't go."
When talking about the fact that the management of some traditional OEMs lacks software thinking and still thinks of “piling up people”, the CEO of a leading autonomous driving company said: “If I hear an OEM say they want to recruit 10,000 software engineers, my first reaction is ‘no chance’.”
Previous article:Autonomous driving middleware part 3: Inventory of active players
Next article:Nvidia's new autonomous driving platform is released, BYD and other car companies will adopt its architecture
- Popular Resources
- Popular amplifiers
- A new chapter in Great Wall Motors R&D: solid-state battery technology leads the future
- Naxin Micro provides full-scenario GaN driver IC solutions
- Interpreting Huawei’s new solid-state battery patent, will it challenge CATL in 2030?
- Are pure electric/plug-in hybrid vehicles going crazy? A Chinese company has launched the world's first -40℃ dischargeable hybrid battery that is not afraid of cold
- How much do you know about intelligent driving domain control: low-end and mid-end models are accelerating their introduction, with integrated driving and parking solutions accounting for the majority
- Foresight Launches Six Advanced Stereo Sensor Suite to Revolutionize Industrial and Automotive 3D Perception
- OPTIMA launches new ORANGETOP QH6 lithium battery to adapt to extreme temperature conditions
- Allegro MicroSystems Introduces Advanced Magnetic and Inductive Position Sensing Solutions
- TDK launches second generation 6-axis IMU for automotive safety applications
- LED chemical incompatibility test to see which chemicals LEDs can be used with
- Application of ARM9 hardware coprocessor on WinCE embedded motherboard
- What are the key points for selecting rotor flowmeter?
- LM317 high power charger circuit
- A brief analysis of Embest's application and development of embedded medical devices
- Single-phase RC protection circuit
- stm32 PVD programmable voltage monitor
- Introduction and measurement of edge trigger and level trigger of 51 single chip microcomputer
- Improved design of Linux system software shell protection technology
- What to do if the ABB robot protection device stops
- Analysis of the application of several common contact parts in high-voltage connectors of new energy vehicles
- Wiring harness durability test and contact voltage drop test method
- From probes to power supplies, Tektronix is leading the way in comprehensive innovation in power electronics testing
- From probes to power supplies, Tektronix is leading the way in comprehensive innovation in power electronics testing
- Sn-doped CuO nanostructure-based ethanol gas sensor for real-time drunk driving detection in vehicles
- Design considerations for automotive battery wiring harness
- Do you know all the various motors commonly used in automotive electronics?
- What are the functions of the Internet of Vehicles? What are the uses and benefits of the Internet of Vehicles?
- Power Inverter - A critical safety system for electric vehicles
- Analysis of the information security mechanism of AUTOSAR, the automotive embedded software framework
- The product cannot be connected to the Internet, and the displayed time cannot be updated through the network. The LCD displays the date and time inaccurately. From which angles can it be solved...
- EEWORLD University Hall ---- Top 10 Raspberry Pi Designs in 2019
- Industrial control system software design based on finite state machine.pdf
- Overview of air conditioning automatic control system 2
- See how many PA posts have been viewed in the forum
- 【PLC based on IoT】---Testing RPI-400 performance based on AI
- Smart car based on ESP32 road sign recognition
- Regarding the issue of isolated communication:
- Please recommend a cheap and available microcontroller
- Seven lines of code to implement an ultrasonic rangefinder (Oled screen display)