The chip is a key factor in determining the performance of the smart cockpit, and companies with higher computing power AI chips will shine in the competition.
"At present, the penetration rate of new cars at the level of cabin intelligent configuration in the Chinese market is about 48.8%, and it is expected to exceed 75% by 2025." Recently, IHS Markit, an internationally renowned consulting organization, pointed out in the "White Paper on the Research of Intelligent Cockpit Market and Technology Development Trend".
In fact, in recent years, with the accelerated development of the "new four modernizations" of automobiles represented by "electrification, networking, intelligence, and sharing", automobiles are changing from a "single means of transportation" to a "third space" that integrates leisure, entertainment, office and other functions.
As a space to carry and realize a series of intelligent applications and services for automobiles, the importance of smart cockpits in the development of intelligent vehicles is becoming increasingly apparent.
At the same time, based on the intelligent car as a platform, comprehensive perception outside and inside the car, the intelligent cockpit that integrates speech and touch screen, emotion recognition, gesture recognition, face recognition, location positioning, etc., is becoming the standard of the current intelligent car.
The smart cockpit ushered in a period of market explosion
According to an IHS Markit survey, in the consideration of user car purchase, the configuration level of cockpit intelligent technology has become the second largest category of key elements after safety configuration, and its importance has exceeded the key elements of traditional car purchase such as power, space and price.
It is worth noting that compared with many mature markets, Chinese consumers who rely on mobile Internet and smart phones have paid more attention to the technological configuration of smart cockpits and play a more important role in car purchase decisions.
With the continuous improvement of consumer demand, creating a "differentiated, personalized, emotional, simplified and social" car cockpit with "active + intelligent" human-computer interaction has become the development trend of future intelligent cockpit.
IHS Markit expects that by 2030, the global market size of automotive smart cockpits will reach 68.1 billion US dollars, and by then, the domestic market size will also exceed 160 billion yuan. China's share of the global market will rise from the current 23% to about 37%, making it the world's leading smart cockpit market.
The huge market space has also attracted many giants to enter the game. Among the participants in the intelligent cockpit ecology, in addition to vehicle companies, Internet companies such as Huawei, Qualcomm, Baidu, and Toyota appear frequently. Players are trying to "get their hands on" in every aspect of the smart cockpit, including dashboards, central control screens, head-up displays (HUDs), audio hardware and software, telematics equipment, and cockpit domain controllers.
However, despite the large number of players, in the face of consumers' higher demand for smart cockpits, chip and algorithm research and development enterprises have begun to become core participants in smart cockpits and ecosystems, which will provide strong technical support and computing power support for various scenario services in the "third space".
"The future of smart cars is a computer on four wheels, which also means that the core component system of the car will produce subversive changes, the most core device is the car intelligent chip." Horizon founder and CEO Yu Kai has said publicly many times.
In fact, from the initial buttons, touch screen, to the current voice, gesture interaction, and then to the future of multi-mode interaction, emotional interaction, the car cockpit is gradually becoming more intelligent and more active, the realization of these functions depends not only on the breakthrough of interactive technology, but also on the research and development and popularization of hardware sensors and chip computing power behind it.
In the face of the market with unlimited prospects for smart cockpits, the smart chip companies that are the first to invest in this market will have a first-mover advantage and are expected to grow into the leader and shaper of the smart cockpit market with partners. Among them, as the only enterprise in China to achieve mass production of automotive intelligent chip front-loading, Horizon continues to occupy the market share. By the end of 2020, the front loading and shipment of Horizon Journey chips has exceeded 160,000 pieces.
High-computing chips drive industry change
Different from the intelligent driver assistance system, the main components of the intelligent cockpit include the in-vehicle infotainment system, instrument panel, head-up display (HUD), streaming media rearview mirror, voice control and so on.
From 2000 to the present, for more than 20 years, with the continuous expansion of users' functional needs for car cockpits, intelligent cockpits have undergone the evolution from safety warning, route navigation, voice recognition to integrated intelligent and networked technologies, software and hardware, continuous learning and iteration into a life complex with perception and decision-making capabilities.
In this process, the traditional cockpit chip has also been continuously upgraded from supporting a single "security" demand to an AI intelligent chip that takes into account multiple service needs and higher computing power.
On the one hand, with the increasing complexity of the intelligent cockpit function, the centralized electronic and electrical architecture has begun to gradually replace the distributed electronic and electrical architecture, and several high-computing chips are replacing dozens or even hundreds of computing power electronic control units in the past car.
On the other hand, with the upgrading of the size of sensors and interaction modes in the cockpit, it is more necessary to rely on the computing power of the cockpit system chip. Chip computing power and structural design greatly affect the functions of the intelligent cockpit, taking the central control multimedia, voice as an example, the chip computing power will affect the response time of the central control multimedia to the user's needs, and at the same time affect the intelligent cockpit's recognition accuracy of the user's voice, recognition speed and other parameters.
Taking the perception algorithm as an example, with the development of cockpit field functions from DMS to IMS, the number and resolution of cameras will continue to upgrade, and more sensors will be introduced, and these innovations will put forward higher computing power requirements for AI chips carrying cockpit computer vision technology. For example, the influence of the number of cameras on the detection of the whole map is linear, and in the case of other impact factors remaining unchanged, two cameras are 2 times the computing power required for one camera. For the same algorithm, multiple impact factors work together, which will lead to a significant increase in the requirements for hash rate.
In addition, the increase in chip computing power makes it easier to achieve rapid development, which is of great significance to OEMs and users. Hundreds of perception algorithms create personalized upper-layer applications for OEMs and complete differentiated strategic deployments; computing power support provides software iteration space for the vehicle life cycle, and can meet the perception function needs of thousands of people.
This demand for chip computing power has also made chip and algorithm R&D enterprises gradually become core participants in the intelligent ecosystem, such as Horizon, NXP, Ambarella, SenseTime, iFLYTEK, etc.
Taking Horizon as an example, facing the smart cockpit market, at this year's Shanghai Auto Show, Horizon launched the Horizon Halo™ in-vehicle intelligent interaction solution, and equipped with a series of vehicle-grade AI chips with different computing power needs, which is also the first intelligent cockpit solution based on domestic AI chip front-loading mass production.
An open ecosystem reshapes the industrial landscape
There is no doubt that the smart cockpit industry is becoming the next outlet for rapid development after smart phones, global technology giants such as Qualcomm, Intel, Visteon, Samsung, Panasonic, Denso, etc. are concentrating in this field, Baidu, Ali, Tencent, Huawei, iFLYTEK, Si Bi Chi and other Internet technology giants have entered the game, hoping to reshape the industrial pattern and dominate the development of the industry.
However, exactly who can stand out in this 100 billion-dollar market is still uncertain. However, it is certain that the industry and service ecology of the smart cockpit is more complex, and it is a multi-party ecosystem, with participants including OEMs, algorithm and chip companies, Internet companies, ICT companies, Tier1 suppliers and governments.
At the same time, the car cockpit has entered a high degree of intelligence, and the future ecological cooperation mode will become inevitable. With the development of technology and the continuous extension of the cockpit industry chain, the future ecosystem cooperation model will focus on "ecological collaboration" and "cross-border extension".
Among the ecological participants, as the ultimate integrator, automakers need to concentrate the roles of software and hardware, functions and ecological service providers to complete the full ecological delivery from vehicle manufacturing to travel services, explore new supply chain cooperation models with algorithms and chip companies, cooperate with Internet companies to explore different travel and life scenarios, and improve the initiative and accuracy of services based on data analysis to create a cockpit service ecology.
The intelligent cockpit chip suppliers and algorithm research and development enterprises represented by horizons will provide technical support and computing power support for various scenario services in the "third space", and rely on the continuous evolution of related product technical capabilities to improve user experience and enrich the ecosystem of intelligent cockpits.
At present, chip companies such as Horizon and NXP have begun in-depth cooperation with many car companies to build an open and win-win intelligent car core ecology.
Taking Horizon as an example, Horizon Halo™, an intelligent cockpit human-computer interaction computing platform equipped with the Journey 2 chip, has achieved mass production on Chang'an's main model, UNI-T, which can realize innovative and active interactive functions such as line of sight tracking, graded fatigue detection, multi-mode lip recognition, driver behavior recognition, intelligent emotion capture and gesture recognition. Horizon's new generation of high-performance vehicle-grade AI chip Journey 3 is also undergoing mass production cooperation with a number of car companies and Tier 1 companies in the field of smart cockpits.
It is worth noting that while continuing to meet the commercial landing of smart cars in the current stage, Horizon will also launch the Journey 5 series chips with AI hashrate of up to 128 TOPS and the Halo 5.0 solution based on Journey 5 this year.
In the past decade, with the development of technology and the improvement of chip computing power, all people's imagination of smart cockpits is becoming a reality little by little; and for the future, consumers' demand for automatic driving, multi-mode interaction, and emotional interaction will be more intense, and the safer, smarter and more active car cockpits rely on breakthroughs in interaction technology, which is behind the research and development and popularization of hardware sensors and chip computing power.
Driven by technologies such as perception systems, algorithms, and AI chips, especially the research and development and popularization of large-scale computing chips, it will also bring unlimited possibilities to the development of intelligent cockpits.