In recent years, breakthroughs have been made in intelligent driving technology, from the popularization of assisted driving functions to the gradual maturity of autonomous driving technology, which not only brings people a more convenient and efficient travel experience, but also triggers a deep transformation of the automotive industry chain, becoming a strategic highland for many car companies and technology companies to compete fiercely. Huawei, as a world-renowned technology giant, has entered the intelligent driving market with its accumulated technical background in the fields of communications, chips, and artificial intelligence. Since its foray into this field, Huawei's ADS advanced intelligent driving system has attracted much attention, and its continuous iteration and upgrading of the version has demonstrated its strong strength in technological innovation. From the early ADS 1.0 version to build the basic framework, to the key technological breakthroughs in ADS 2.0, to the cutting-edge technological changes brought about by the latest ADS 3.0, Huawei's intelligent driving solution has continuously optimized the hardware configuration and algorithm architecture, and continuously expanded and deepened its functions, rapidly emerging in the intelligent driving market and becoming an important force in promoting the development of the industry.
According to Huajin Securities' intelligent driving series report: A brief analysis of Huawei's intelligent driving solution, it can be concluded that Huawei's ADS intelligent driving solution in the field of intelligent driving has distinctive characteristics, and it has always adhered to the multi-sensor fusion route of combining lidar, millimeter-wave radar, and camera. With the prominent demand for cost reduction in the industry, the solution has made adjustments in hardware configuration, gradually changing from overconfiguration to meeting actual needs, effectively reducing the overall hardware cost. Specifically, the number of sensors in a single vehicle has decreased, such as the number of lidars has been reduced from 3 to 1, the number of millimeter-wave radars has been reduced from 6R to 3R, and the number of cameras has also decreased. The computing power on the side of the car has been reduced from 400TOPS to 200TOPS, and the "Max + Pro" dual-version intelligent driving hardware configuration mode is being explored.
The "Max + Pro" dual-version intelligent driving hardware configuration mode refers to whether it can support urban NOA (Navigate on Autopilot) as the boundary, of which the Max version supports urban NOA and is equipped with high computing power, which is ready for future intelligent driving system upgrades; The Pro version supports high-speed NOA and has a relatively low computing power configuration, ensuring the realization of the more mature L2 intelligent driving.
For example, Huawei explored this model on the Zhijie S7 released in November 2023, and its Max series (including Max, Max+, Max RS, and Ultra versions) supports urban NCA, comes standard with one 192-line LiDAR, and has a suggested retail price of 269,800 yuan and above; The Pro version supports high-speed NCA, no lidar, and has a suggested retail price of 249,800 yuan.
Figure: Huawei's intelligent driving solution market analysis report (source: HIMA official website).
From the perspective of the industry, since 2022, some mainstream smart driving manufacturers, including Xpeng and Ideal, have also adopted this model in hardware configuration. For example, Xpeng G9, which was officially released by Xpeng in late September 2022, has launched a Max version that supports XNGP and a Pro version that supports Xpilot.
In terms of algorithm architecture, Huawei ADS has been continuously upgraded. From ADS 1.0 in 2021 to ADS 2.0 in 2023, obstacle recognition has changed from manual annotation to autonomous decision-making, and road recognition has also shifted from a mapped scheme that relies on high-precision maps to a non-mapped scheme. The ADS 3.0 version, launched in April 2024, adopts an end-to-end large model, and its architecture is GOD+PDP with hierarchical perception and decision-making, which is different from Tesla's "big unification" model. In terms of perception and decision-making planning, Huawei's ADS continues to be optimized, not only improving the recognition ability of alien shapes and rare obstacles through the GOD algorithm, but also realizing real-time lane recognition and path planning without a map with the help of the RCR algorithm, accelerating the generalization speed and update frequency of intelligent driving solutions in cities.
Huawei's intelligent driving ecosystem is constantly expanding, and it has launched in-depth cooperation with five major car companies, and continues to expand the scope of cooperation with end car companies through three cooperation models. The sales scale of its ADS high-end package may have exceeded 3 billion yuan, and the car BU is about to break even. In addition, Huawei is actively involved in the formulation of domestic intelligent driving standards, and is expected to continue to lead the development of China's intelligent driving industry in the future with its own technological advantages. The continuous improvement of Huawei's intelligent driving technology has also made the competitiveness of its cooperative models increasingly enhanced.
