Lidar’s role in the autonomous vehicle race
By: Harrison Rubin
Over the past several years, there has been significant progress made in the race towards autonomous vehicles. Almost every single major automobile company is diligently working on their own unique car which they hope will be the first to market. While the concept of autonomous cars is fascinating, many enthusiasts do not understand how the technology even works. One of the most integral components of autonomization is lidar. In fact, lidar is so important to autonomous vehicles that every manufacturer developing an autonomous car (except for Tesla) has incorporated it into their vehicle.
So the big question is why is there all this hype around lidar?
Light detection and ranging (lidar) is a device similar to a radar that emits laser pulses which are then processed to develop a 3D image of the world around it. While it sounds complicated, the process of how lidar develops a 3D map is really quite simple. The system first emits laser pulses that travel outward until they make contact with an object. When they hit an object, they reflect back to the system, which the system then uses to measure the distance between itself and the object. It then creates a 3D image of the world around it based on the data it receives from the laser emissions. The data created by lidar is also extremely easy to interpret since it is merely measures and records the distances between objects. After the data has been processed, software identifies the objects that surround the vehicle and determines all the possible movements the object could make.
The reason why lidar has become such an essential part of the autonomous vehicle race is because it can map the roads significantly better than any other device, even better than cameras. One major flaw in cameras is that they require software to convert the recorded 2D images into 3D, which has been an extremely difficult feat. Lidar already processes the objects in 3D, which saves time and money for autonomous car manufacturers. Additionally, lidar emits hundreds of thousands of pulses every second, which provides a real-time display of what is occurring around the vehicle. Another problem with cameras is that they do not work well in harsh weather conditions, such as rain or snow.
There are only two challenges facing lidar: its production cost and durability. Lidar is extremely expensive to produce, costing at the very least $4,000. However, most lidar systems used in autonomous cars cost significantly more than this. The reason for this is that these cars need a system that is powerful. The second problem is that lidar systems might not be able to handle the harsh conditions of the road. Roads are fraught with potholes and other conditions, and it is unclear how well these systems will work in these conditions. There are also additional concerns around general wear and tear due to harsh weather conditions.
Lidar is an extremely exciting component of autonomous vehicles that will undoubtedly play an integral role. It will be interesting to see how the systems improve as companies continue on their quest towards producing an autonomous vehicle.
___________________________________________________________________
https://www.wired.com/story/lidar-self-driving-cars-luminar-video/
https://www.wired.com/2014/09/velodyne-lidar-self-driving-cars/
https://www.ff.com/us/futuresight/what-is-lidar/
https://www.youtube.com/watch?v=ibIzthKKZWY
https://www.wired.com/2016/10/teslas-self-driving-car-plan-seems-insane-just-might-work/
