A Blue Ocean - The Prospect of Automotive Software (I)

With the increasing application of technology in in-vehicle terminals, in-vehicle software will surely become the core of the differentiation of intelligent vehicles. Throughout the process, in-vehicle software architectures are also evolving – gradually transitioning to SOA. Compared with traditional cars, automotive intelligence can provide a rich and perceptible driving experience (which is unmatched by traditional cars), which is the core of the differentiation of today's intelligent cars. As the core of vehicle intelligence, the value of on-board software will increase rapidly in the future.

"Hardware embedding, software upgrade" has become the mainstream strategy of mainstream OEMs, the next five years will become a key node in the development of autonomous driving Level-3 (perhaps higher level), for car companies with leading software and algorithm capabilities, software suppliers can get important opportunities in this process (this is also a number of Internet manufacturers Xiaomi, Juchang, DJI, Baidu are laying out the basic on-board software and automatic driving algorithm iteration).

Under the torrent of on-board intelligence and networking, the electronic and electrical architecture (E/E) has gradually changed to a domain architecture and a central computing architecture (I personally believe that this trend must be divided for a long time, and finally it is believed that the domain controller and the central calculator process data together), and the hardware architecture is the hardware foundation for realizing the software-defined car. In order to cope with different needs, the development of in-vehicle software and functions needs to be rapidly iterated, and the Internet-like SOA (service-oriented software architecture) has gradually been recognized by the industry (mainstream car companies are establishing their own software research institutes in sole proprietorships or joint ventures). In soA architecture, the service component interfaces (APIs) required for functions are initially defined as standardized, and the architectural function deployment of software no longer depends on specific hardware platforms, operating systems, etc. Under this architecture, it is lightly coupled, flexible and easy to expand, and with the AUTOSAR in-vehicle software architecture, the "soft and hard separation" of the car is truly realized.

For intelligent car in-vehicle software architecture can be divided into:

- > system software layer;

- > functional software layer;

- > application software layer.

1. The basis of the vehicle operating system (OS) is the system software level, and the stable operating system is the basis for the realization of all functions. The operating system can provide efficient and stable environmental support for Application and Function, and the operating system dispatches the underlying hardware resources according to the task priority (the role of the bridge is the core function of on-board intelligence).

2. Functional software (Functional Software) includes the core functional module of automatic driving and the corresponding middleware, using the public function module, Developer can save time and resources, and focus on developing the business level function of automatic driving.

3. Application Software includes intelligent car scene algorithms (automatic parking, different application scenarios of automatic driving), intelligent cockpit functions (better reflecting the experience of intelligent cars to drivers), massive data (algorithm iterative optimization), high-precision maps (scene-based rendering comparison) and other content, which is the core of the on-board intelligent cockpit and automatic driving solutions that reflect the differences of each technology.

In such a big environment (national policy support, massive influx of funds), the trend of software-defined cars has been created. Manufacturers with full-stack software and hardware capabilities will reflect their corresponding value in this process. This wave will also have a considerable impact on the current OEM-Supplier pattern.

Traditional Tier1, OEM to withdraw part of the system function development rights is the trend of the times. Tier1 in order to avoid becoming an OEM hardware oem (starting from itself, how much initiative needs to be occupied in the whole process), the future trend is that more and more Tier1 is committed to creating a full-stack technical capability of "hardware + underlying software + middleware + application software algorithm + system integration" (Bosch, Huawei, etc.), and more players are constantly pouring in (DJI, zero run, etc.). As the next commanding height of market share, everyone is making efforts in the layout.

For pure software suppliers, OEMs have begun to consciously bypass the traditional Tier1 and seek direct cooperation with software vendors (interesting here, either the OEM and the software vendor joint venture to establish an OEM software research institute, or the sole proprietorship to set up a software research center, digging people and digging teams).

Under the general trend of software-defined automobiles, the software architecture of intelligent vehicles slowly reflects the characteristics of the previous Internet SOA. In-vehicle software has become the core of the differentiation of intelligent vehicles.

Under the trend of vehicle intelligence, "software-defined cars" have become the consensus of the automotive industry (and also the topic that has been very popular in recent years). Software Defined Vehicles-SDV refers to software's deep involvement in the whole process of car definition, development, verification, sales, and service. And constantly change and optimize each process, to achieve continuous optimization of experience, continuous optimization of process, continuous creation of value. The whole process is also a rapid iterative rhythm, which is the same as the previous Internet software architecture.

Based on the current situation of on-board materials, each year OEMs for the launch of different models in the hardware configuration gradually converged, whether foreign or domestic, OEMs in the hardware field has experienced a long period of competition, on-board hardware and its cost of continuous improvement of space is limited. Automotive intelligence can provide drivers with rich, perceptible value and a new driving experience (the most direct is perhaps the most effective selling point), so in-vehicle software intelligence is the focus of the differentiation of each OEM new model. On-board stability software and autonomous driving algorithms have become the core points of the competition of car companies, and the technical threshold of car manufacturing has also been transformed: tens of thousands of parts and components are combined and integrated capabilities - > hundreds of millions of lines of code integration optimization capabilities (robustness of code operation).

For OEMs, in the field of intelligent competition, the robustness of on-board software, the computing power of on-board chips, and the optimization of automatic driving algorithms have become the core technology points of intelligent vehicles in the new era. The new four modernizations (intelligent, networked, electrified, and shared) have become the general trend of the transformation of the automobile industry, and the on-board products have gradually changed from traditional transportation machinery tools to a new generation of intelligent terminals with perception and decision-making capabilities (from the previous transportation tools to the "CPU" of walking and running). As Wang Chuanfu said:

"Electrification is the first half, and intelligence is the second half. Intelligence is a greater change, and the ecology created is beyond imagination.

In the era of electric intelligent vehicles, the core technologies of the three electrics such as on-board power batteries, electric drives, and battery optimization algorithms are the core of the power domain competition. Compared with traditional cars, the nature of intelligent electric vehicles has changed, computing and software have become the core of the car, and software capabilities will also be the strategic commanding heights of the vehicle function.

As software is the key to realizing the new four modernizations of automobiles:

Networking

-> deploy new functions through OTA functions, faster and more efficient updates, and achieve rapid iteration of in-vehicle software; - > some functions of future cars can be operated through the cloud; - > integrate third-party related services.

electrification

-> based on new needs and functional advancements, new electronic components are introduced in vehicles; - > reduce the consumption of power batteries (energy) through advanced software algorithms;

automation

-> the trend of increasing sensors and actuators in the car, the calculation is calculated through the domain controller or the central processor; -> corresponding functions have higher requirements for computing power and in-vehicle communication capabilities, and the traditional CAN bus cooperates with the trend of vehicle Ethernet; - > the increase in the proportion of on-board software, and has higher requirements for the robustness of on-board software.

Sharing

-> call shared travel services through application software (mobile APP); - > for drivers, customized driving experience (one of the purpose points of Didi Car).

"Hardware embedding, software upgrade" as the mainstream strategy at this stage, the automatic intelligent driving system of passenger cars as a whole is at the L3 and below level (a considerable part of the level), and the status quo of intelligent driving technology is still in the process of continuous iteration. According to past experience, in-vehicle products have a long life cycle (5-10 years), so the upper limit of computing power of the in-vehicle computing platform (for the calculation and processing of massive data transmitted by Sensor, there are higher requirements for computing power) determines the upper limit of software service upgrades that can be carried during the life cycle of the vehicle. In order to ensure the continuous software upgrade capability of the vehicle in the whole life cycle, the OEMs adopt the strategy of "hardware presetting, software upgrade" in intelligent driving, first using high-power car specification chip pre-embedding, and then continuous software update (OTA strategy) after the car.

In view of the policy impact, the mainland intelligent car development speed is ahead of the world (excluding Tesla, the entire national level of the mainland is in the lead), and the landing of autonomous driving L3 and above is expected to bring greater opportunities to the industry (and it is a greater opportunity for our personnel). Driven by huge consumer demand, we can build an ecological environment of benign interaction and promote the rapid improvement of the level of industry standards (overtaking in curves).

For you and me to work as engineers in the automotive industry, it is also a once-in-a-lifetime opportunity. Standing at such an outlet, you need to calm down and grasp the core of the outlet - on-board software, whether it is the evolution of electronic and electrical architecture, automatic driving upgrades, grasp the core, improve their core competitiveness, and will be invincible. For yourself, based on the on-board diagnosis, step by step to broaden your own face, for the future, you also have some visions:

- > 1. Based on new requirements and functional advancements, especially changes in electrical and electronic architecture, the introduction of domain controllers and central processing units, how to define and implement functions for on-board diagnostics;

- >2, autonomous driving will introduce multi-core processors, but also introduce more functional modules (perception fusion, high-precision maps, etc.), how to define and deploy diagnostic functions between modules;

->3, OTA scenario, what are the requirements for vehicle information security, how to achieve? Combined with the Internet security policy, it is actually applied to the vehicle operating environment, and what specific scenarios need to be improved;

- >4, the increase in the proportion of software, functional safety at the software level, and functional degradation strategy;

The above are all based on their own points of interest, want to broaden the points, if you can also cause you to think before the text, fortunately !!!

-----------------------------------

Author Bio | A man in slippers

Automotive Electronics Engineer

Public number: on-board diagnostic technology

Come, improve a little bit every day!