buffer don't have space! when running with pcap file input about hesailidar_general_ros HOT 7 OPEN

this is happening randomly on the same pcap. sometimes the pcap works sometimes it doesnt

from hesailidar_general_ros.

Hi tianshiz,

Please go through our readme for launching steps.
It seems like you didn't give correct lidar type, eg. "lidar_type:="Pandar64""

Thanks,

from hesailidar_general_ros.

Hi Fred,

Sorry I didn't update the post. I did eventually add the lidar_type:="Pandar64". But, even running w/ the sample data provided by Hesai for Pandar64, I've found the application to be unreliable. Sometimes it will work, sometimes it will give me this error.

Are you not seeing that w/ this data?

from hesailidar_general_ros.

Hi Fred,
I have a model XT32M2X
and used the following to run the pcap file. The problem is that its inconsistent at best.
Sometimes its able to run through the entire pcap, but other times, it could be playing halfway through and then I get the "buffer don't have space!" warning, and then the pcap stops playing. 50% of the time, it gives me the warning immediately and I have to restart.

I would have to re-run this several times and get "lucky" for it to work. Can I please get some support here?

roslaunch hesai_lidar hesai_lidar.launch pcap_file="test.pcap" lidar_type:="PandarXTM" frame_id="PandarXTM" lidar_correction_file:="correction.csv"

test.pcap: https://drive.google.com/file/d/1ADX8B5q9BwdAubOGtbT0M1psw-vWYXFZ/view?usp=sharing

correction.csv:
`

from hesailidar_general_ros.

Hi tianshiz,

I downloaded the pcap file you provided and run roslaunch on my PC, monitored the log ouputs and point clouds on RIVZ.
I repeated this test for 10 times, there are no error message and the whole pcap file can be prcessed well.
Here are possible reasons I can think of about the issue on your end:

1. The PC you are using might be not fast enough, is it possible to change a better PC to test?
2. Is ther any other tasks added to the ROS driver code? The issue might be caused by additional tasks added

Thanks,
Fred

from hesailidar_general_ros.

Thank you for taking the time to test that out.
I am actually using a amazon aws cloud desktop with the following specs. I run a essentially barebones vanilla ros noetic docker image on it to run this driver.

the desktop uses Intel Xeon E5-2676 v3 (Haswell) processors that run at a base clock frequency of 2.4 GHz and can deliver clock speeds as high as 3.0 GHz with Intel Turbo Boost. While monitoring my CPU and MEM usage, neither of them are anywhere close to maxing out while running the driver.

As for your #2 question, the only thing that I do ontop of running this process is that I also start a rosbag to bag these pointcloud topics. But I've actually seen this issue without bagging as well.

from hesailidar_general_ros.

Hi Tianshi,

This is highly posssible caused by the running environment, you can try use a local PC to test.

Thanks
Fred

from hesailidar_general_ros.