Large-scale applications and stable, mass production of advanced driver-assistance systems (ADAS) are key to accelerating the democratization of high-level driving technologies and benefiting a broader range of consumers, said Zhang Lin, chairman and CEO of domestic ADAS tech company Freetech.

"With the accelerated mass rollout of Level-2 assisted driving products, more advanced functions will gradually reach mid and lower-end models, further promoting equal access to intelligent driving," Zhang said.

For example, its city Navigate on Autopilot (NOA) solution entered mass production in July and is now installed in nearly eight vehicle models, he added.

"Technology inclusiveness is further reinforced through scaled mass production, supported by high delivery efficiency and cost performance, forming a virtuous cycle of research and development input, complete product portfolio, mass delivery, and economies of scale," Zhang said.

Zhang's remarks came after the company's 5 millionth ADAS product — an urban environment NOA controller — rolled off the line last week.

"From the launch of our intelligent manufacturing base in 2022, to surpassing 1 million units in 2023, and now reaching 5 million units, we expect total production to exceed 10 million units in the future," said Zhang.

The senior executive attributed this large-scale delivery and effective cost control to the company's self-developed integrated software-hardware platform, ODIN 3.0.

"The core of this approach is to transform capabilities traditionally improved only through hardware iteration into software modules that can continuously evolve on a unified platform," Zhang said.

For example, although the evolution of millimeter-wave radar appears to be a hardware upgrade, at the platform level, it becomes a modular building block, with new AI algorithms, higher resolution, enhanced detection range and signal-to-noise ratio ultimately abstracted into software capabilities that feed back into system upgrades, enabling mass production data to more efficiently support product iteration, he explained.

"This means hardware is no longer an isolated device, but part of an integrated platform that evolves together with perception, decision-making and cloud services," Zhang said.

In addition, R&D methods may also fundamentally change under this architecture, he added.

"Instead of restarting from scratch, the system evolves along a stable backbone. When one module requires an upgrade, only that module needs to be updated, avoiding unnecessary disruption to proven components," Zhang said.

As the platform matures and scales up, these advantages extend from development to commercialization.

"On the software side, modular reuse reduces labor and time costs; on the hardware side, mass manufacturing improves procurement and production efficiency — once the production process is established, it can be rapidly replicated across lines," Zhang said.

"As volume expands, cooperation willingness and bargaining power among upstream and downstream suppliers also increase, further reducing hardware costs and thereby broadening its applicability," he added.