Track Hyper | Qualcomm Snapdragon Ride: Breaking the Deadlock in ADAS Popularization

Author: Zhou Yuan / Wall Street Insight

As cars transition from mere transportation tools to intelligent mobile terminals, the prevalence of Advanced Driver Assistance Systems (ADAS) is becoming an important benchmark for measuring the level of intelligence in the automotive industry.

Qualcomm's technical white paper titled "Snapdragon Ride: Driving the Adoption of ADAS Among Chinese Automakers and Consumers," released in late June, reveals the key role of technological architecture innovation and ecosystem collaboration in the process of ADAS moving from high-end configurations to mass adoption.

As the world's largest automobile producer and consumer, China's automobile production is expected to exceed 31 million units in 2024 (data source: China Association of Automobile Manufacturers), providing a vast application scenario for ADAS technology.

However, this fertile ground also conceals challenges: the high-density traffic on urban roads, complex driving behaviors, and the dynamic changes in map updates and regulations all impose higher demands on the adaptability of ADAS systems.

Chinese consumer demand also shows significant stratification characteristics: users in first- and second-tier cities have a strong willingness to explore advanced autonomous driving features, while the lower-tier markets focus more on basic safety assistance functions.

How to meet the advanced technological requirements while controlling costs to adapt to different market demands has become the key for companies to break through.

Taking basic functions such as Automatic Emergency Braking (AEB) and Lane Keeping Assistance (LKA) as examples, these technologies have been proven to effectively reduce traffic accident rates, but there are significant differences in penetration rates across different consumer tiers, reflecting the complexity of market demand.

In addition, China's unique traffic environment poses special requirements for ADAS technology.

For instance, the rapid growth of electric vehicle ownership has led to a demand for charging scenario management, and the characteristics of frequent interactions between non-motor vehicles and pedestrians in mixed traffic flows require ADAS systems to possess stronger scene recognition and dynamic decision-making capabilities.

The "impossible triangle" of performance, power consumption, and cost has long plagued the industry in the development of ADAS technology.

The Snapdragon Ride platform seeks to find a balance among these three through underlying architecture innovation: the heterogeneous computing architecture uses a 4nm process technology, deeply integrating CPU, GPU, and NPU, with different computing units performing their respective roles, ensuring the ability to process complex algorithms while effectively controlling energy consumption.

For example, when processing multi-sensor data fusion tasks, the NPU can be dedicated to accelerating AI algorithms, significantly improving computational efficiency compared to traditional general-purpose processors.

The cabin and driving integration design is a rethinking of traditional automotive electronic and electrical architecture.

Snapdragon Ride Flex integrates cabin entertainment and driving assistance functions into a single SoC, not only reducing hardware costs but also minimizing system response delays through optimized data links.

The advantages brought by this design are reflected in multiple aspects: on one hand, it reduces data transmission losses between different chips; on the other hand, a unified software architecture facilitates functional expansion and system upgrades, providing automakers with a more flexible development model.

At the perception level, the Qualcomm platform supports a multi-modal input configuration of 11 cameras and 7 radars, combined with map-free ADAS technology, allowing vehicles to break free from reliance on high-precision maps This perception of "self-reliance" constructs decision-making models through real-time environmental perception, aligning more closely with the actual needs of the dynamic changes in China's road environment.

In scenarios where map data cannot be updated in a timely manner, such as temporary traffic control and road construction, the no-map ADAS system can still operate normally, ensuring driving safety.

The large-scale implementation of technology relies on the support of an ecosystem. The Snapdragon Ride platform has been validated in over 60 countries and regions worldwide, accumulating 482 million kilometers of testing mileage, and collaborating with more than 20 automotive companies to build a vast technology implementation ecosystem.

To this end, Qualcomm not only provides hardware and software but also offers full-process support for automotive companies "from research and development to mass production" through cloud management environments and data simulation factories.

The standardized development environment significantly lowers the development threshold for ADAS technology. Automotive companies do not need to build a development platform from scratch and can directly develop functions based on the hardware development boards and software stacks provided by Qualcomm, shortening the R&D cycle.

The shared data resources provide a foundation for algorithm optimization, and the inclusion of partners like Momenta and Desay SV provides further enrichment of technical application scenarios, forming a virtuous cycle of "technology supply - scenario validation - optimization iteration."

In terms of safety assurance, it features built-in ASIL-D level functional safety microcontrollers, with redundancy mechanisms, safe booting, and other characteristics, certified by third parties. This comprehensive safety design ensures the reliability of the ADAS system in complex application scenarios.

The popularization of ADAS is essentially a technological equality movement and a reflection of the automotive industry's transformation from hardware determinism to software-defined vehicles.

Through architectural innovation, Qualcomm has reduced corresponding hardware costs, and by collaborating within the ecosystem, it has shortened development cycles, aiming to make advanced driving assistance technologies accessible to more consumers.

From the perspective of industry competition, the ADAS market has formed a multi-party competitive landscape. Traditional chip manufacturers, tech giants, and emerging startups are all entering the fray, with increasingly significant differences in technology route selection and function definition.

Tesla insists on a pure vision solution, while some automotive companies adopt a lidar + multi-sensor fusion route. This divergence in technological routes reflects different judgments within the industry regarding future development directions and exacerbates the complexity of market competition.

As the dependence of ADAS systems on data deepens, how to ensure user data privacy, prevent data leakage, and how to establish reasonable decision-making ethics in emergencies all require the industry to jointly explore solutions.

The practice of Qualcomm's Snapdragon Ride platform provides a feasible path for the popularization of ADAS. However, the ultimate value of technology lies in whether it can effectively improve travel safety and enhance quality of life.

In the continuous iteration of technology and ecology, ADAS will eventually become a standard configuration in the era of intelligent mobility, and the participants in this transformation are reshaping the future landscape of the automotive industry