LiDAR Technology in Autonomous Vehicles
Almost every major automotive manufacturer around the world agree that LiDAR will be the primary sensing technology that is used in self-driving vehicles—and LiDAR has answered the challenge by delivering continuously improving performance and reliability, coupled with dropping price points.
So, what's next on the horizon for LiDAR and autonomous vehicles?
I recently visited with Louay Eldada, CEO of Quanergy, which produces LiDAR and smart sensing solutions.This is what Eldada had to say:
What makes LiDAR the de facto sensing technology of the auto industry?
Eldada: There are over 50 different automakers around the world. All but one (Tesla, which uses a combination of radar, cameras and ultrasound technology) have endorsed LiDAR as their primary sensing technology for autonomous vehicles.
What are the challenges of incorporating LiDAR technology into autonomous vehicles?
Eldada: There are significant challenges when it comes to using LiDAR on autonomous vehicles because of all of the things that LiDAR is expected to deliver. These elements include high performance, high reliability and low cost. The performance LiDAR must deliver includes having sufficient range to detect what is going on in a street scene, accuracy of detection, high resolution, and the ability to operate in diverse sets of weather conditions.
What recent improvements has LiDAR achieved?
Eldada: We have achieved reliability and technology performance for LiDAR, while also reducing price so that a single LiDAR unit can be acquired for under $1,000 and drop to as low as $200 per unit for large volume purchases.
What future LiDAR technology advances do you foresee?
Eldada: LiDAR will continue to improve its capabilities while also lowering the price points it’s able to deliver. But one of the breakthroughs we have achieved with LiDAR is the ability to use solid state technology that can reduce the size of a LiDAR device to the size of a deck of cards, with a sensing range that has no minimum range limits. The competing technology that is based on ultra-sonic sensors has a minimum range limit of 1- to 2-meters, so we have eliminated that “blind spot.”
What is the next challenge that LiDAR on autonomous vehicles faces?
Eldada: The automobile industry needs to converge on a standard for optical equipment such as LiDAR that is deployed on vehicles. Right now, the practice is to employ five LiDAR units on a vehicle: one in the front, two on each side, and together, they are used for perception, localization and navigation. What this combination of devices is delivering is what is known as a level 4 autonomous optimal system, where humans do not have to interfere with the autonomous functioning of the vehicles.
Is level 4 the highest level of vehicle autonomy?
Eldada: There actually is a level 5 level of vehicle autonomy that represents total automation with no need for human involvement whatsoever. It differs from level 4 autonomy in that level 4 allows for human overrides and participation with the automation, such as the ability to take over control of a vehicle’s steering by grabbing the steering wheel.
There was a debate between Google and the State of California, with California insisting that autonomous vehicles come equipped with steering wheels and Google insisting that steering wheels weren't needed. What happened in that debate?
Eldada: What many onlookers didn’t know was that the Google vehicle in question was already equipped with retractable brakes and a retractable steering wheel. You simply push a button in the car and they both come out, so there was a failover mechanism in place that would allow a human to manually regain control of the vehicle. At this point, it’s hard to know if we will ever achieve level 5 automation in autonomous vehicles on public streets, as they are limited for the foreseeable future to geofenced areas.
What about consumer pushback against autonomous vehicles? For instance, aren’t there a lot of people who just enjoy driving?
Eldada: Undoubtedly, there will be people who enjoy driving and would prefer not to have an autonomous vehicle do that for them. I am one of them. I enjoy getting out with my vehicle and driving it. However, what I don't enjoy is having to do a 10-hour drive to a distant destination or having to endure rush hour bumper-to-bumper traffic. Both of these situations tend to produce a lot of stress. I think what you'll see is drivers making selective choices about when they want to drive and when they would prefer the vehicle do it for them.
The original version of this article, written by Mary E. Shacklett, appeared on GeoDataPoint, an AVT sister BNP publication site.