Chongqing - Chongqing-based Toramon, a leader in intelligent driving technology, has kicked off mass production of its bionic visual perception car components, set to supply key players in the automotive market.

The company has completed delivery of 3,000 units in Q1 2025, with clients including some top electric vehicle (EV) makers, primarily in commercial vehicle producers.

As intelligent assisted driving evolves, two main approaches dominate: pure vision systems and multi-sensor fusion combining cameras with LiDAR. But Chongqing-based Toramon is carving its own path with innovative bionic visual perception technology, setting itself apart in this competitive field.

Pure vision, championed by Tesla, relies on cameras and AI models for spatial reasoning and decision-making. In contrast, the LiDAR-vision fusion approach boosts accuracy with redundant sensing hardware. However, both face challenges like high computational costs, pricey hardware, and difficulties adapting to new scenarios.

Toramon engineers analyze ways to enhance the use of human-eye-inspired technology in predicting vehicle collision risks. (Photo/Zhang Jinhui)

“We believe driving doesn’t require full cognitive engagement. Humans rely primarily on their eyes, and instinctively react to danger without fully recognizing what the object is,” said Luo Rui, Brand Director of Toramon.

Human vision has two pathways: one for recognizing objects, the other for sensing motion, allowing quick reactions to potential collisions. Inspired by this, Toramon’s system combines traditional AI object recognition with a motion perception algorithm. It detects collision risks from movement, responding instantly without the need for repeated high-precision training, cutting down on data annotation and reducing costs.

The cost advantage makes Toramon’s technology ideal for commercial vehicles, including freight vehicles and passenger vehicles, which are more price-sensitive yet offer vast market potential. Unlike the crowded passenger car market, commercial vehicles are better suited for scalable deployment.

Currently, 80% of Toramon’s devices are targeted at this segment. “While new-energy carmakers spend tens of millions monthly on AI training, we achieve similar results with just millions,” Luo said.

Toramon testers are calibrating the multifunctional all-in-one smart in-vehicle terminal and other equipment installed on test vehicles. (Photo/Zhang Jinhui)

For now, Toramon mainly uses the technology to focus on L2-level forward driver assistance, tailored to typical commercial vehicle use cases regulated under Chinese law, such as identifying two-wheelers, the rear of lead vehicles, and pedestrians. Looking ahead, the company plans to expand the capabilities of its system toward higher-level intelligent assisted driving.