On April 11, Ford released a video of “Project Nightonomy,” its nighttime autonomous vehicle tests. The green-tinted night vision video shows a prototype Ford Fusion navigating the company’s Arizona test track while only using LIDAR. The Fusion handles the test with ease, navigating the dark desert night as effortlessly as it handles a well-lit road.
“As soon as I look out the window all I saw was blackness,” the safety driver in the front seat states in the video.
It’s a blackness that is becoming more and more rare in the United States. Urban sprawl and street lamps along highways have lit a lot of the shadows that used to cover the country. Light pollution makes stargazing practically impossible, and, as a Florida Atlantic University study notes, not a single species evolved for continuous lighting, and not a single species “truly benefits from it in the long run.” Birds, reptiles, insects, and even humans are negatively affected by the never ending glow of modern life.
If Project Nightonomy is a sign of the future, perhaps autonomous cars and their LIDAR — Light Detection And Ranging — systems are a solution to the light pollution problem. Ford says it will have the technology to safely put self-driving cars on the road by 2020. Will street lights start to disappear with their arrival?
According to Michael Clamann, senior research scientist at Duke University’s Humans and Autonomy Lab, the answer is no.
“The whole point of having multiple systems on these cars is that they have redundancy,” Clamann tells Inverse. “There are certain things that LIDAR is good at, there are certain things that cameras are good at, and there are certain things that radar is good at.”
Autonomous vehicles are currently being developed to use all three systems together. Where one system is weak, another system takes over. Cameras, obviously, struggle in the dark. The speed and accuracy of radar decrease the further away objects are from the sensors. LIDAR, which was almost prohibitively expensive until recently, also can’t see through rain or snow.
Project Nightonomy proved an autonomous vehicle could operate while only using LIDAR on a wide open road with zero to few obstacles. Small environmental variations could quickly escalate into huge problems for self-driving cars if backup systems weren’t able to take over.
LIDAR maps the surrounding area with pulsing infrared light and a photodetector. Light is shot out, the beam reflects off of objects, and then is picked back up by the LIDAR system. The distance is determined by calculating the time between the two signals.
“By shooting off thousands of pulses in a well-designed pattern,” the Texas Instrument’s blog explains, “you can map the surrounding area of a vehicle to distinguish cars, people, trees and other obstacles.”
When snow and heavy rain get in the way, LIDAR sensors lose their ability to “see” their surroundings.
That’s not to say that autonomous vehicles are incapable of driving during the night. They will, however need the same light that humans need.
“I think unless there’s a major jump in technology, relying on a single system without redundancy wouldn’t be safe,” Clamann says. “If you look at the way a lot of these systems work, by combining all these things, these cars can drive at night. It’s just that you wouldn’t want to drive solely with LIDAR.”