Sorting out key technologies of smart cockpit software and hardware

The car cockpit has evolved from the original mechanical control to electronic control, and then developed into the current intelligent operation. The intelligent cockpit is a complete system composed of different cockpit electronics, and its key technologies are mainly composed of four parts. The first part is mechanical technology, including variable body technology and interior structure technology. It is a trend that cars can be retracted and folded according to different modes in the future. The cockpit needs to be able to continuously adjust and change the interior space according to the passenger's use needs in different scenarios. The second part is electronic hardware technology, including: display hardware (screen, HUD, etc.), interactive devices, cameras, radio frequency , communication units/gateways, cockpit domain controllers , chips , etc. The third part is software technology, including: operating system , basic software, virtualization technology, artificial intelligence (user portrait, situational awareness, multimodal fusion interaction, etc.), application development ( mainly Android ), instrument software development (mainly QNX), TBOX software development ( Linux ), cloud services (big data, information security, etc.), protocol stacks. The fourth part is two major supporting technologies, namely artificial intelligence technology and cloud computing technology. The accuracy of future intelligent algorithms determines the differentiation of different brands of intelligent cockpits and is the key to influencing the in-car experience. Next, we will briefly introduce the use and development of key software and hardware technologies in the cockpit field.

operating system

There are currently five mainstream smart cockpit operating systems: Linux, Android, QNX, WinCE, and Hongmeng OS. In recent years, the entertainment and information service attributes of smart cockpits have become increasingly prominent, and operating systems have become the key to automakers' intelligent networking layout and mastery of core technologies.

After entering the 5G era, whether it is the Internet of Vehicles or connected cars, intelligent operating systems will become the core of future cars.

The current market has the following two characteristics: 1. The era of software-defined cars has arrived; 2. More and more technology companies and traditional car companies have increased their investment and layout in the field of basic automotive software. This cross-border trend will intensify. As core participants in smart cars, companies have increased their R&D investment and built stronger software capabilities to ensure that they continue to have a place in future competition.

Virtual Machines

Hypervisor , also known as VMM (virtual machine monitor), has become an indispensable software system for cockpit electronics as the performance of cockpit processors becomes stronger, cockpit screens become more and more, and cockpit electronics cover more and more functions.

Mainstream virtualization software review: Common virtual machines include BlackBerry's QNX, Intel- led ACRN , Mob ic a-represented XEN, Panasonic- acquired Open Synergy's COQOS, Green Hills' Integrity, Japan's eSOL, SYSGO's Pi ke, Mentor 's Nucleus, Samsung Harman's Redbend, and EPAM's Xen.

Domestic virtual machines are also on the rise (such as Zhongling Zhixing). QNX is recognized as the most mature and secure cockpit virtual machine operating system. QNX is the most widely used, but the fees are relatively high, including entry fees, seat fees, service fees and licensing fees (charged by screen volume).

Multimodal interaction

Smart car human-computer interaction technologies include buttons/knobs, touch, voice interaction, gesture recognition, biometrics, etc., and multi-modal fusion interaction has become a trend. At this stage, the human-computer interaction in the car is mostly a combination of touch, voice, biometrics, AR technology, etc., integrated into the display system, smart seats, interior and other areas in the car cabin to improve the efficiency and convenience of human-computer interaction. The overall market situation of human-computer interaction: the models and models with central control screens have the highest installation rate and have entered the mature stage; the growth potential of LCD instruments and HUDs in the later period is great; the model installation rate of voice recognition systems has reached 55.5%; biometrics (such as face recognition, health monitoring) are mostly installed in high-end models and need further penetration; these interaction methods have their own advantages and disadvantages, and it is difficult to completely replace each other, so multi-modal interaction is an inevitable trend of development.

HMI Design

Human-computer interaction involves software, mainly including the design side and the implementation side:

Design side: Automotive HMI design is divided into UX and UI. The software involved in the design process mainly includes Photoshop, Effects , Illustrator, 3Dmax , etc. Different OEM design departments, Tier 1, and design service companies will have certain differences;

Implementation side: mainly includes operating system and HMI integration software.

High-precision map

High-precision positioning combination under the conditions of Internet of Vehicles = high-precision map + GNSS + IMU. Vehicle positioning solutions are diverse depending on the scenario and positioning performance requirements. In most Internet of Vehicles application scenarios, accurate positioning is usually achieved through the integration of multiple technologies, including GNSS, inertial measurement unit (IMU), sensors , and high-precision maps; GNSS is the most basic positioning method.

Compared with ordinary maps, high-precision maps provide map information with higher accuracy and richer content, mainly serving autonomous driving. Currently, L2+ and above autonomous driving solutions generally have a clear dependence on high-precision maps. As shown in the figure below, high-precision maps can be said to be the upstream of the data flow.

Vehicle-road-cloud collaborative Internet of Vehicles

The Internet of Vehicles can be divided into two concepts: broad and narrow. The broad Internet of Vehicles includes V2X vehicle-road collaboration (safety needs) + Telematics (entertainment needs). The narrow Internet of Vehicles refers to V2X vehicle-road collaboration, which mainly includes three major areas: V2V (vehicle to vehicle), V2I (vehicle to road), and V2P (vehicle to people). The technologies involved in the Internet of Vehicles mainly include four parts: cloud platform, V2X, OTA, and information security. The accelerated expansion of intelligent driving penetration rate and the growth of car ownership drive the expansion of the Internet of Vehicles market space: China's road mileage and car ownership are still on the rise.

Domain control chip

With the popularity of domain control and central computing platform architecture, and the explosion of demand for intelligent car products , the concept of software-defined cars has been implemented. Domain control chips that are strongly related to smart cockpits are developing in the following aspects: the computing power of chips is getting stronger; the GPU capabilities and video processing capabilities of chips are getting stronger; chips integrate AI- related IP cores; chips are developing towards a higher level of functional safety; cockpits and autonomous driving domain control are showing a trend of integration, and chips are developing towards central computing platforms; domestic domain control chips are catching up.