There are generally two camps when it comes to the topic of autonomous vehicle sensing:
Camp Fusion believes that fusing multiple sensor types—vision, LiDAR, radar, etc.—is a prerequisite for safety. Members of this camp, such as Zoox and the current iteration of Waymo, argue that sensors like LiDAR and radar provide superhuman perception capabilities that cameras alone cannot match.
On the other side, Camp Vision argues that high-resolution cameras, coupled with massive training datasets and inference compute, offer a simplicity and cost structure that beats the “extrasensory” advantages of the fusion approach. Proponents like Tesla and Wayve are betting that a camera-first strategy is the scalable path forward.
However, one criticism dogs the camera-only approach: cameras need to see.
Raindrops, snow, ice, and road grime can blind a camera in seconds. This necessitates constant cleaning, a challenge that has sparked a sub-industry of its own. As Viodi has covered previously, companies like Marelli have integrated cleaning nozzles into their sensor suites, while Murata has detailed its own ambitions with an innovative integrated camera cleaning solution.
Atomize the Rain Drop or Snow Flake #
Enter Maxell with a compact new contender that leverages the physics of sound to keep the robot’s eyes clear.
As seen in the above video, Maxell’s Ultrasonic Cleaning Lens Unit (UCLU) uses ultrasonic waves to atomize raindrops the instant they hit the lens surface, preventing them from sticking and distorting the image. Crucially, the device handles winter weather without a traditional heater. Instead, it simply shifts the frequency band of the ultrasonic vibrations to generate thermal energy, melting snow and ice using the same underlying mechanism.
While the product is currently listed as “under development,” Maxell’s press materials indicate that mass production is targeted for late 2026. Perhaps most impressively, the device boasts a small form factor and extremely low power consumption—less than 5 Watts even when in snow-melting mode.
As the industry races toward Level 4 and 5 autonomy, this type of “set it and forget it” hardware could be the missing link that ensures an autonomous vehicle’s vision is always clear and ready to sense the road ahead.
[Note: the above text was directed and edited by the author, but written by Google’s Gemini].
Leave a Reply