Things you need to know about LiDAR: how to choose a lidar?

2023-03-21

Lidar has many tags as a "new species" in the automotive industry: FOV, reflectivity, ROI, and more.

While the industry hasn’t developed a unified standard, how can we identify a high-performance lidar?

Li Auto L9 is equipped with Hesai’s AT128 lidar

Today, we will share with you some tips for choosing a lidar.

Can it see far distances?

It is easy to understand that the farther lidar can see, the more sufficient reaction time is given for the intelligent assisted driving system to make decisions.

The "detection distance" of lidar usually refers to its "ranging capability," not the "maximum range." There are differences between the two concepts.

"Maximum range" is the farthest detection distance without restrictions, while "ranging capability" is the farthest detection distance under standard working conditions. And one of the essential standards is the 10% reflectivity target.

Next, we would like to introduce a new concept: reflectivity, the degree to which an object reflects light.

Given the same material, white objects have a higher reflectivity than black objects, smooth objects have higher reflectivity than rough objects, and objects with higher reflectivity are easier to perceive.

For example, a black tire has a typical 10% reflectivity target.

Hesai AT128 has a range of 200m@10% reflectivity

Under normal circumstances, the "maximum range" of the same lidar is greater than the "ranging capability." For example, a lidar brand that advertises a maximum detection distance of up to 400 meters may only have a ranging capability of 200 meters when the reflectivity is limited to 10%.

"Ranging capability" is a more meaningful indicator for assisted driving. Therefore, it is not enough to see far; it must be able to recognize effectively. This brings us to the next tip, whether it can see clearly.

Can it see clearly?

More is needed for a high-performance lidar to see far and clearly.

For example, in the picture below, there are only a few points, and you can not identify what the target is. Only when the resolution is high enough can it be effectively determined by the system.

Simulated point cloud

So what is the core indicator that represents the "resolution"?

For a camera, a "pixel" is the core indicator of resolution. The denser the pixel "dots", the higher the resolution.

Lidar, a perception sensor, can be seen as a three-dimensional camera. The denser the three-dimensional pixel "points," the higher the resolution. This three-dimensional pixel point is the number of point clouds generated by the lidar per second, the point frequency.

Lidar’s resolution of 1.53 million points/1 million points/750,000 points per second

A higher point frequency creates a higher resolution, thus providing a clearer field of view and bringing safety to the assisted driving system.

Is it stable and reliable?

As the "eye" of smart cars, stability and reliability are the prerequisites for lidar to ensure safety.

First off, a qualified automotive-grade lidar needs to go through dozens of rigorous reliability tests, including high and low-temperature tests to withstand harsh weather, temperature cycle vibration tests to withstand long-term vibration, and mechanical shock to withstand instantaneous shock tests.

There are various environmental tests such as ultraviolet aging, dustproof, waterproof, salt spray, ice water impact, etc., to ensure the lidar can work under harsh conditions.

Therefore, people who buy a car with lidar do not have to worry about lidar damage caused by car washing, high temperatures, or bumps during driving.

Lidar reliability test

On the other hand, whether the lidar design is "robust" enough will also affect its lifespan. For example, its scanning method, the frequency of mechanical movement, and the number of lasers are factors that will all affect its reliability.

Copyright © 2024 Hesai Group. All Rights Reserved.

crosschevron-right