Smart car: the powerful computing terminal of the future

As the core domains of automobile intelligence, two things have extended the role of cars from travel tools to “mobile third space” – the smart cockpit and automated driving. This is the subject of a recent report published by Beijing-based firm Minsheng Securities, a company that provides securities brokerage, securities underwriting and other services. In this article, we summarize some of the report’s key findings.

The traditional supply chain landscape will transform in future as the value driver of the automobile industry shifts from hardware to software. The supply chain will develop from its pre-existing linear structure to a flat, reticulated form centering on OEMs, whose role will change – from manufacturers to service providers, seeking out the high-profit opportunities at different points along the supply chain. The future competition of OEMs will focus on the intelligent performance of their smart cars and their ability to provide continuous service for customers.

Automobiles will gradually be transformed from primarily being mechanical vehicles into being seen as electronic products with mechanical functions, as their performance shifts from hardware-dominated to software-hardware co-dominant. According to McKinsey’s forecast, by 2030, the proportion of software and hardware in the global auto market will reach 30% and 41% respectively. The proportion of software has nearly tripled from 2016.

Smart cockpit: starting with the Internet, embracing the Metaverse

The smart cockpit’s evolution is marked by configuration upgrades and architecture integration, which will take place through innovation, iteration and subversion, and move toward the Metaverse concept.

• Configuration innovation: the combination of LCD instrumentation with a central control screen, HUD, ambient lighting, seats and the acoustic system creates in-car “scenes” and multi-modal interaction. Meanwhile, networking equipment such as T-BOX connects various AIoT devices, establishing the smart automobile as a cloud-connected part of the wider Internet of Things.
• Architecture iteration: in terms of hardware architecture, cockpit chips will focus on smart and networking functions; the domain controller is gaining importance particularly in terms of engineering, hardware adaptation, and software capabilities. In software architecture, the ultimate competition will take place at the ecosystem level: the operating system, algorithms and HMI (Human–Machine Interface) will be the keys to the competitive edge of a given ecosystem. Operating system design will have to provide the features that consumers demand addressing the often-contradictory characteristics of openness and security; cockpit algorithms will focus more on offering multi-modal interaction to users; HMI provides a differentiated experience of human-computer interaction.
• Ecosystem subversion: the smart cockpit will serve as a portal to access AioT terminals such as smart home products, smart phones and wearables, and become a conduit for Metaverse technologies.  

A Metaverse meeting from inside the car (image source: Minsheng Securities Research Institute) 

Autonomous driving: the deconstruction of “perception-decision-execution” technologies

Autonomous driving equips the car with an “invisible driver” and makes it a “mobile third space” with the design of the smart cockpit. Autonomous driving technology can be sub-divided into three elements: perception, decision-making, and execution.

• Perception:
  • Camera: functionality will shift to focusing on image acquisition and may usher in a golden age with rising output and lower price. According to data from analysts ICV, China’s vehicle camera shipments will increase from 16.9 million units in 2017 to 42.63 million in 2020.
  • Lidar: According to the official statement of key OEMs, most of their models equipped with lidar are expected to be in mass production in 2021–2022. The lidar ADAS (Advanced Driver Assistance Systems) segment will grow rapidly, with a CAGR of 94% from 2021 to 2026. According to Yole, the overall market size of lidar is expected to reach US$5.7 billion during the same period, with a CAGR of 23%. Camera image acquisition vs lidar (image sources: Cheyun, Minsheng Securities Research Institute)
• High-precision maps: commercial use of high-precision maps may reshape the industry structure.
• Decision-making:
  • Hardware architecture: the autonomous driving chip and domain controller industry is booming, but there is competition in vehicle models and price differentiation.
  • Software architecture: unlike cockpit software, autonomous driving software focuses more on functional iteration, so its operating system needs to ensure security and low latency; the algorithm will also shift to a data-driven "functional algorithm".
  • Execution: upgraded to X-by-wire technology dominated by active control.

In conclusion, this article provides just an overview of the ways in which industry insiders and experts expect the smart car – a potentially epoch-defining product of the Internet era – to develop in the coming years. Self-driving and smart cockpit features will be changing the fabric of the industry around the world, sooner rather than later. Whether you are on the cusp of stepping into the automotive industry, or already an auto-sector veteran, these are trends and developments that you can’t afford to ignore.