A Look At Lidar, The 54-Year Old Technology Fueling Google And Uber's Self-Driving Ambitions
What Is Lidar?
Created shortly after the invention of lasers, lidar works in a manner similar to sonar. It bounces light off of distant objects in order to determine their shape and distance. Given the fact that light travels at a uniform speed, the time it takes for the light to bounce off an object and then return can be used to determine how close the object is to the car.
Of course, this would be a problem for oncoming traffic were it not for the fact that lidar systems use light that falls within the near-infrared spectrum, which means they're invisible to humans.
Beyond merely telling distance, lidar can also be used to analyze the type of object that is hit by the light, providing data on whether it is dark or light or, more relevant to the roadways, hard or soft. Many lidar systems are capable of handling millions of measurements per second. This allows the computer to paint an accurate picture of the car's surroundings.
Working Out The Kinks
Most of the major players in the self-driving car industry rely on some variant of lidar systems. Unfortunately, lidar is also a very expensive technology, with some models costing as much as $8,000, which is simply too much for the mass market.
Tesla has tried to get around this by avoiding lidar altogether. Instead, Elon Musk's cars will rely on a combination of cameras and radar. Of course, if this were the ideal solution then everyone would do it. Unfortunately, cameras don't work well at night whereas lidar actually functions better at night than during the day because of the lack of interference from other light sources.
That's not to say that costs are the only hindrance to lidar systems. The current models bear a resemblance to emergency sirens and aren't something that most people would want on their cars. Additionally, poor weather conditions, particularly rain, can reduce lidar's effectiveness.
A Silver Bullet?
Because of its high costs, lidar isn't likely to be the silver bullet that self-driving car companies are hoping for, but it may very well be part of the solution.
"Every kind of sensor is making progress, which is a good thing," said roboticist Chistoph Mertz of Carnegie Mellon University. "We are going to need robust, redundant systems."
Eric Brackett Tech Times editor Eric Brackett is a tech junkie and a gamer, covering science and technology. Follow him on Facebook and Twitter for updates and his random thoughts on the latest trends in gaming, tech, and comic books.