Sensing technology is the powerful tool that gives vision to next-generation automobiles.
Equipped with a device array anchored by LiDAR and video cameras, the average advanced vehicle of today “sees” and “feels” the road and surrounding topography more clearly than its predecessors ever could. However, there is one major problem with these devices – like our own eyes, their effectiveness can be badly affected by the weather. Rain, fog, mud, almost any challenge from Mother Nature can cloud those eyes, threatening the effectiveness of the advanced assistive features that depend on them.
“Cameras, LiDARs, IR cameras are all susceptible to bad weather. It’s basic physics,” said Miles Flamenbaum, CEO of Actasys. “If something is on the lens, like a raindrop, the incoming visual data (images for a camera, laser beams for a LiDAR) is refracted and the sensor may not be able to receive enough data to inform the underlying ADAS system.”
There is much to clean, however. Tim Sylvester, CEO and CTO of Integrated Roadways, pointed out that: “It’s not just weather that can affect functionality but sunlight at the wrong angle, glare, glinting off steel or chrome, something hanging from the rearview mirror, an errant stone, ice formation, a tree branch falling on the roof.”
Some in the connected vehicle space argue that certain devices are more at the mercy of the weather than others. Specialist LiDAR company Velodyne’s vice-president of government affairs Christina Aizcorbe, said about cameras: “Common driving conditions like bright sunlight, glare, rain, and darkness can present real challenges. As a result, cameras are more susceptible to unpredictable blind spots and to generating false positives and/or negatives. Where cameras are dependent on ambient light conditions, LiDAR provides its own light source, allowing it to work in all lighting and weather conditions.”
So, under threat of the elements, what’s a device maker or automaker to do? There are various approaches that can be taken to mitigate or even eliminate the problem of inclement weather. Straightforward cleaning of camera and LiDAR lenses is an obvious and direct approach, although there are challenges with this.
“It’s generally accepted that approaches to sensor cleaning that [touch] or interfere in any way with the sensor are not acceptable – this eliminates things such as small wipers, or anything that interferes with sensor data capture,” said Flamenbaum. “Secondly, the industry is interested in deploying a sensor cleaning system that can address the widest range of weather in environmental conditions.”
Others in the industry propose radical solutions to the challenges thrown at us by our environment. Sylvester thinks the best answer for these and others that bedevil current ADAS solutions, lies outside of the vehicle entirely. “There are countless problems with on-vehicle technology that will never be solved if we assume we don’t need a network,” he said.
Sylvester’s vision is of a robust one built directly into the roads and highways. This kind of always-awake, digital road brain, he believes, could eliminate the challenges of device vision-obscuring entirely. “With the network in the road, we don’t need cameras or LiDAR because the road can feel exactly where every vehicle is, without any of the dangerous, expensive computerized guesswork,” he said. “Instead of spending all our computer power constantly remapping an area with cameras or LiDAR, we can take the fixed, unchanging map of the area, overlay it with the real-time vehicle positions, and disseminate that map to every vehicle that can receive it.” That concept is rather attractive and promising, although there are a host of challenges we’ll have to overcome to even begin contemplating a massive, always on, in-road digital network.
Concentrating on the here and the now, we must figure out workable, and hopefully inexpensive and easily integrable, solutions to keep those vehicle “eyes” open and fully clear. After all, “every sensor (camera, LiDAR, etc.) that is mounted or placed outside of the car, and is exposed to the elements of nature is susceptible to data loss in inclement weather,” Flamenbaum says. “Each of these sensor types experiences data loss differently under a range of conditions.”
