Lidar giving wrong distance

asked 2022-02-10 06:06:43 -0500

Roshan gravatar image

updated 2022-02-11 04:36:13 -0500

Hello, I have a physical Turtlebot3 set up behind an object. The distance between the LiDAR scanner and the object is 1.16 meters, but the LiDAR measures at angle 0 measures between 1.11 and 1.15 meters, mostly around 1.13 and 1.14. It seems the LiDAR is consistently measuring a few cms less than what it should. Is there any way to calibrate the LiDAR?

There also seems to be a lot of noise in the measurements, watching the lidar detections on Rviz some of the points jump around quite a lot. Is the physical LiDAR just more noisy and less accurate than in the simulations?

Additional information: It's a TB3 Waffle Pi, the LiDAR is a LDS-01

edit retag flag offensive close merge delete


Is the LiDAR position on the TB3 taken into the consideration? With the calculated correct offsets and so on. I think your comments about a lot of noise in physical measurements are also correct

ljaniec gravatar image ljaniec  ( 2022-02-11 04:09:18 -0500 )edit

Hello, yes the offset between the base_scan and base_link is taken into account. I've been printing out both distances when doing this just to compare and see if there isn't anything wrong with my transform but the transform does not seem to be the problem.

I did take a closer look at the individual measurements of each of the laser scan, and they seemed a little off since there were big differences (up to 10 cm) between one laser and the ones next to it, and the same laser measurements were also jumping up to 10cm from one time instance to the next.

In the simulations most of the measurements seem to be more accurate and within the normal error that are given in the LiDAR specifications.

Roshan gravatar image Roshan  ( 2022-02-11 04:19:49 -0500 )edit

Were these laser working at the same time? What about the reflecting surfaces? I found this list of LiDAR error sources/reasons: Maybe it could be helpful (whole LiDAR Tutorial too:

ljaniec gravatar image ljaniec  ( 2022-02-11 04:39:52 -0500 )edit

No the lasers were not working at the same time, the LiDAR sensor spins around, but shouldn't be affected that much if the two measurements you compare are next to each other. The object that should be detected has an aruco marker attached to the back of it covering most of the object, and I did consider that the black marker would cause issues, but covering the marker up with some cardboard did not affect the results too much. Could maybe be that the marker is not perfectly 90 degrees on the back of the object, but it looks close. Will have to test more

Roshan gravatar image Roshan  ( 2022-02-11 04:49:16 -0500 )edit