Automakers Diverge on Side-Sensing Approaches: LiDAR, Pure Vision, and UWB Coexist

In advanced driver-assistance systems (ADAS), forward perception has long centered on LiDAR performance, yet lateral perception capability is equally critical. Recently unveiled at China’s Ministry of Industry and Information Technology (MIIT), the high-end variants of the Zunjie V800 and S800 feature cameras or perception modules mounted on the B-pillars, while the all-new Aito M9 and Zhiji V9 add cameras to their C-pillars. These elevated-position cameras complement those installed on fenders and side mirrors, working in tandem with solid-state LiDARs on the fenders and ultrasonic sensors on the front and rear bumper sides to accurately detect low-profile and small lateral obstacles—addressing blind-spot challenges common in large vehicles. Currently, mainstream flagship models generally adopt a combination of fender-mounted and side-mirror cameras, solid-state LiDARs, and ultrasonic radar. However, automakers are clearly diverging in their technical approaches: XPeng adheres to a pure vision strategy, relying on two cameras on the side mirrors, one on the fender, and ultrasonic sensors near the wheel arches for lateral perception. Its flagship GX model hasn’t significantly upgraded lateral hardware but instead focuses on building user trust through refined driving experiences. Tesla pursues an even more aggressive pure-vision approach, eliminating all ultrasonic sensors entirely. For lateral perception, it depends on fender- and B-pillar-mounted cameras, with the latter offering a higher and wider field of view toward the front-side area—though this requires structural modifications to the body and poses sealing challenges. Li Auto continues its multi-sensor fusion philosophy, equipping both fenders and side mirrors with cameras alongside solid-state LiDARs, while replacing ultrasonic radar with UWB (Ultra-Wideband) technology, which it claims enables more precise detection of suspended objects and allows for cleaner exterior styling. Xiaomi blends Tesla’s approach with mainstream domestic solutions, installing cameras on the fenders, mirror mounts, and B-pillars to cover the rear-lateral, low blind zones, and front-lateral areas respectively. Overall, Huawei and Li Auto emphasize redundancy and multimodal sensor fusion; Tesla and XPeng prioritize optimization of pure-vision algorithms; and Xiaomi attempts a hybrid, comprehensive layout. As automakers race toward full autonomy, these divergent hardware configurations fundamentally reflect differing philosophies regarding the optimal synergy between software and hardware.