An article explains the current status of automotive operating systems

The Software Branch of the China Association of Automobile Manufacturers (hereinafter referred to as the Software Branch) released the China Automotive Operating System Open Source Plan (hereinafter referred to as the Open Source Plan). Puhua Basic Software and Software Branch member units FAW, Dongfeng, Changan, China Auto Innovation, China Electronics Technology Group Corporation 32nd Institute, Western Intelligent Link, Horizon Robotics, CoreDrive, Advanced Operating System Innovation Center, University of Electronic Science and Technology of China and other enterprises and universities formed an open source co-construction partner to implement the open source plan. The release of this open source plan marks that the Chinese automobile industry is moving towards a higher level of openness in the development model of intelligent connected vehicles.

Preface: In the context of software-defined cars, the operating system is the soul of the development of the automotive ecosystem. With the development of electric, intelligent, and connected cars, the automotive operating system has become one of the most important components of the vehicle, which to a certain extent determines the safety, comfort, intelligence level, and overall performance of the vehicle.

1. Introduction to automotive operating systems

The automotive operating system is a real-time security platform software running on a heterogeneous distributed hardware architecture. It provides a functional framework for vehicle and component perception, planning, and control, and supports a collection of software for the intelligent connected driving ecosystem. It is an important foundation and core support for the safe, real-time, and efficient operation of the automotive intelligent computing basic platform.

Automotive operating systems include safe in-vehicle operating systems, intelligent driving operating systems and intelligent cockpit operating systems.

① Safe in-vehicle operating system

Safe in-vehicle operating systems are mainly aimed at classic vehicle control areas, such as power systems, chassis systems, and body systems. This type of operating system has extremely high requirements for real-time and security, and its ecological development has become mature.

The safe vehicle operating system is mainly a real-time operating system RTOS, and its main application object is ECU. The most basic requirement of ECU for the safe vehicle operating system is high real-time performance. The system needs to complete specified actions such as resource allocation and task synchronization within a specified time. Embedded real-time operating systems have the advantages of high reliability, real-time performance, interactivity and multi-channel performance. The system response is extremely high, usually at the millisecond or microsecond level, meeting the high real-time requirements.

At present, mainstream safe in-vehicle operating systems are compatible with two types of automotive electronic software standards: OSEK/VDX and Classic AUTOSAR. Among them, the Classic platform is based on the OSEK/VDX standard and defines the technical specifications of safe in-vehicle operating systems.

② Intelligent driving operating system

With the development of intelligent and networked technologies, the perception fusion, decision-making planning and control execution functions of smart cars have brought more complex algorithms and generated a large amount of data, requiring higher computing power and data communication capabilities. The secure in-vehicle operating system based on OSEK/VDX and Classic AUTOSAR software architecture can no longer meet the development needs of future autonomous vehicles. The AUTOSAR organization has launched the Adaptive AUTOSAR platform for centralized electronic and electrical architectures for more complex domain controllers and central computing platforms.

Adaptive AUTOSAR defines an operating system based on the POSIX standard, which can provide standardized platform interfaces and application services for operating systems that support the POSIX standard and different application requirements, mainly to meet the development needs of automotive intelligence. Adaptive AUTOSAR is in its early stages of development, and it will take some time for its ecosystem construction to be widely recognized by Tier1 and OEMs.

Smart driving operating systems are mainly used in the field of smart driving and are applied to smart driving domain controllers. This type of operating system has high requirements for security and reliability, as well as high requirements for performance and computing power. This type of operating system is becoming increasingly mature worldwide, but the ecosystem is not yet complete.

③ Intelligent cockpit operating system

The intelligent cockpit operating system mainly provides a control platform for automotive infotainment services and in-vehicle human-computer interaction. It is the operating environment for the car to realize cockpit intelligence and multi-source information fusion. It does not have strict requirements on the real-time and reliability of the operating system.

Smart cockpits of mainstream models mainly include QNX, Linux, Android, etc. QNX occupies the vast majority of the traditional smart cockpit operating systems. In recent years, the entertainment and information service attributes of smart cockpits have become more prominent. Open source Linux and Android, which has a large number of mature information service resources on the mobile phone side, are favored by many OEMs and have become rising stars. In addition, a small number of foreign models also use Win CE as a smart cockpit operating system.

As people's demand for vehicles to transform from simple means of transportation to smart mobile terminals, smart cockpit operating systems need to support diverse applications and services and have rich ecological resources.

2. Introduction to mainstream automotive operating system solutions at home and abroad

With the rapid development of autonomous driving technology, the demand for changes in automobile software, especially operating systems, is increasing. OEMs, Tier 1 suppliers, and autonomous driving software and hardware technology solution providers have invested a lot of manpower, material and financial resources in the research and development of automobile operating systems, hoping to gain a foothold in the era of software-defined cars. The following is a brief introduction to the development and application of mainstream automobile operating systems at home and abroad.

① Tesla Autopilot autonomous driving software architecture

As we all know, Tesla is a leader in autonomous driving technology and industrialization. Its advantage lies in its computing platform as the core, self-developed and leading chip hardware, operating system, platform software, etc. Tesla's autonomous driving software architecture is shown in the figure above. The main feature is that its operating system is based on a single Linux kernel, creating a complete set of autonomous driving software solutions, realizing the entire process of perception, positioning, fusion, decision-making, planning and control.

From the public information, we know that the system is tailored based on Ubuntu, and the Linux kernel has been modified for real-time performance. This kernel is also open source on Github. The deep learning framework is based on PyTorch, and the real-time data processing is based on the open source stream processing platform Kafka. It has 48 independent neural networks for multi-dimensional data processing and has powerful OTA upgrade capabilities. Its FSD (Full Self-Driving) computing platform hardware integrates the smart cockpit domain and the autonomous driving domain. The operating system is upgraded through OTA software, making full use of data and cloud computing ecology to create a new model of automotive product value and service.

② Volkswagen centralized software architecture

In order to accelerate the application of autonomous driving technology, Volkswagen has formed a large team to independently develop the automotive operating system vw.OS. vw.OS adopts the service-oriented software architecture based on Adaptive AUTOSAR, among which the centralized software architecture is shown in the figure above.

The main design features of Volkswagen's new generation EE architecture are:

• Using high-performance processors and high-speed networks

• POSIX-compatible kernel (Linux/QNX, etc.) + Adaptive AUTOSAR

• Decoupling application software and I/O functions to reduce the complexity of the entire system and the dependencies between applications

• Develop user functions efficiently and quickly

• Adopt service-oriented communication

③ Huawei MDC intelligent driving computing platform architecture

Huawei MDC (Mobile Data Center) is positioned as a computing platform for intelligent driving. The platform integrates Huawei's more than 30 years of R&D and manufacturing experience in the ICT field. It is based on CPU and AI processor chips, equipped with an intelligent driving OS, compatible with AUTOSAR, and supports smooth evolution from L2 to L5. Combined with a complete supporting tool chain, customers or ecological partners can develop intelligent driving applications for different application scenarios.

The main features of Huawei's MDC intelligent driving computing platform architecture are:

• Provide software and hardware solutions with high degree of decoupling and independent upgradeability. The hardware upgrade route and software upgrade route are independent.

• Good adaptability to mainstream sensors, supports data access of mainstream GNSS, IMU, camera, lidar, millimeter wave radar and other sensors, and supports front fusion of camera and lidar point clouds;

• It is well adaptable to mainstream middle-layer software, compatible with ROS and AUTOSAR, and supports autonomous control of core components (chips, operating system kernel) of common deep learning frameworks such as Caffe and TensorFlow;