As a new species of LIDAR in smart cars, various new keywords are dazzling: FOV, reflectivity, ROI ……
Today, when the industry has not yet formed a unified standard, how can we tell what kind of LIDAR is really excellent?
Today, we’ll share some tips for picking a LIDAR.
Can you see far?
The further the LIDAR can see, the more response time the intelligent driving system can have to make decisions.
When we talk about LIDAR “detection distance”, we are referring to its “distance measurement capability”, not its “maximum range”, there is a difference between these two concepts.
“Maximum range” is the farthest detection distance without any qualification, while “range capability” is the farthest detection distance under a series of standard operating conditions, one of the important criteria being 10% reflectivity of the target.
Here is a new point of knowledge: reflectivity, refers to the degree of reflection of light objects.
In the same material, white has a higher reflectivity than black, smooth objects have a higher reflectivity than rough objects, and the higher the reflectivity, the easier the object is perceived.
For example, a black tire is a typical 10% reflectance target.
In general, the “maximum range” of the same LIDAR is greater than the “long range capability”. For example, a LIDAR that is advertised to have a maximum detection distance of 400 meters may only have a long range capability of 200 meters with a 10% reflectivity limit.
For assisted driving, “long-range capability” is a relatively more meaningful indicator. In layman’s terms, it is not useful just to see far, but also to be able to effectively identify, which leads to the next article, whether to see clearly.
Can you see clearly?
A good LIDAR light to see far is not enough, but also to see clearly, that is, have a high enough resolution.
For example, in this picture below, there are only a few points, you can’t identify what the target is; only if the resolution is high enough, it can be effectively identified by the system.
What is the core metric that represents the “resolution” of LiDAR?
For a camera, “pixel” is the core indicator representing its resolution, the denser the pixel “dot”, the higher the resolution.
The denser the pixel “dots”, the higher the resolution. The same as a sensing sensor, LIDAR can be regarded as a 3D camera, the denser the 3D pixel “dots”, the higher the resolution. This 3D pixel is the number of point clouds generated by the LIDAR per second, which is also known as the point frequency.
The higher the point frequency, the higher the resolution, the clearer the world in the LIDAR field of view, and the safer it will be for assisted driving systems.
Is it stable and reliable?
As the “eyes” of smart cars, stability and reliability are the prerequisites for LIDAR to play all the safety roles.
First of all, a qualified vehicle-grade LIDAR needs to go through dozens of severe reliability tests: including high and low temperature tests to withstand severe weather, temperature cycling vibration tests to withstand long-term vibration, and mechanical shock tests to withstand transient shocks.
There are also UV aging, dustproof, waterproof, salt spray, ice and water impact and other kinds of environmental tests to ensure that the LiDAR can work normally even under harsh conditions.
Therefore, friends who buy cars with lidar do not have to worry about car washing, high temperature, driving bumps, etc. will damage the lidar, the lidar is still very sturdy car specification level.
On the other hand, whether the LIDAR system is designed to be “robust” enough will also affect its lifespan. For example, the way it scans, the frequency of mechanical movement, the number of lasers, all affect its reliability.
