This method, say the researchers, proved "remarkably" successful. With bugs, for example, their tests showed that the air curtain successfully diverted the vast majority of them away from the self-driving sensors.
While this helped reduce the main problem, say the researchers, insects could still break past the air curtain in some situations, so they needed a way to successfully clean the sensors when necessary. So the researchers developed a cleaning system that features next-generation nozzles next to each camera lens that can spray washer fluid as needed to clean the sensors.
The system uses advanced software algorithms that helps the self-driving vehicles determine when a sensor is dirty, and then specifically target dirty camera lenses, efficiently cleaning each one individually without wasting washer fluid on already-clean sensors. After a sensor(s) has been sprayed down, the tiara releases air through a slot which quickly dries the face of the lens.
The researchers tested the effectiveness of the system by driving one of their test vehicles through the Huron-Manistee National Forests in western Michigan too see how the cleaning system reacted to swarms of bugs. The system has also been equipped on the third generation of their self-driving test vehicles, which are now hitting the streets in Detroit, Pittsburgh, Miami-Dade County and Washington, D.C.
"As fun as some of this development may sound," says Krishnan, "these are not features that would simply be nice to have when self-driving vehicles are ready to be deployed; they are critical functions that vehicles must be able to carry out on their own in order for safe deployment to be possible. Just as we must equip self-driving vehicles with the brains to process what’s happening in their environment, we must also equip them with the tools to deal with that environment — no matter what kind of gunk it decides to throw