Open deb0ch opened 5 years ago
Hi @deb0ch,
Could you please let me know how did you figure out that for the outdoor mode, you should update the value
to Stability
from Sensitivity
for the scan_mode
parameter? I couldn't find that anywhere, not even in the source code. On the other hand, the website says
RPLIDRA A3 supports to work under two modes alternatively: enhanced mode and outdoor mode.
Do they have enough documentation for the ROS package? I couldn't find that either!
It's surprising that it's been 1.5 years since you created this issue and they haven't responded yet! Anyway, were you able to figure this out?
Thank you so much in advance!
One more thing is, in rplidar_ar.launch, if I keep scan_mode
to Sensitivity
and launch view_rplidar_a3.launch then in the terminal, I gets this:
current scan mode: Sensitivity, max_distance: 25.0 m, Point number: 15.9K , angle_compensate: 4
On the other hand, if I change scan_mode
to Stability
then it gives this:
current scan mode: Stability, max_distance: 25.0 m, Point number: 10.0K , angle_compensate: 2
So, now I'm even more confused because, in Stability
mode, I'm getting fewer points compared to Sensitivity
mode. So, which scan_mode
should I use for the outdoor application?
I just tried my new RPLidar A3 in my backyard, and even in
Stability
mode the signal was drastically noisier than with the RPLidar A2 under the exact same conditions.I ran them both using
rplidar_ros
on a Raspberry Pi 3 Model B, the A2 using the defaultrplidar.launch
, so inExpress
mode, and the A3 usingrplidar_a3.launch
, changing the mode toStability
.It was clear the the source of the noise was the sun: at the moment the clouds were occulting the sun the signal improved drastically, going back to noise when the sun shined again. The RPLidar A3 wasn't even able to see a big grey garbage can that was 3m in front of it under sunlight but it could with the clouds. A lot of points were falsely detected close to the lidar, at about 0.3m. The A2 worked perfectly under both conditions.
Here is an illustration of the problem with the RPLidar A3:
On this curve, the X axis is time in seconds and the Y axis the distance in meters to an obstacle, a big wooden plank. To get it, I discarded all the points except in a 90º field of view in front of the lidar, and then among those points, selected the one closest to the lidar and applied a
cos
of the angle to get the forward distance. I then moved the lidar in front of the big wooden plank. You can see an area where the signal gets cristal clear: when clouds passed in front of the sun.For comparison, here is the RPLidar A2 under the same conditions and hard sunlight:
The signal is a lot better under the same sunlight, which I did not expect.
Robustness to sunlight is critical in my application. What kind of settings should I use for the RPLidar A3, that would make it resist sunlight better than
Stability
?