Open mitchallain opened 6 years ago
mitchallain
commented
6 years ago This appears to have been due to a reflector that was too far away (est. 10 m). Masking that reflector resolved the issues. I will continue to seek whether this is a sensor or driver issue. SICK technical support seemed unaware of any similar issues.
jrgnicho
commented
6 years ago @MitchAllain thanks for reporting this. I haven't seen this issue before but It seems to me like a hardware problem based on what you described.
mitchallain
commented
6 years ago I can provide a link to a rosbag if there is interest.
mitchallain
commented
5 years ago @jrgnicho What leads you to believe that this may be a hardware issue? After comparing the packets over wireshark while experiencing the issue to the packets when the sensor and driver work normally, there doesn't seem to be any difference in packet length or structure.
I am leaning toward this being a driver issue but I am unable to find any leads in the driver source code. Even after masking the farthest reflector (12 days ago), I occasionally will experience the issue again with a different reflector (always the farthest visible reflector). Masking that reflector returns the NAV350 + driver to its expected output.
Please let me know if I should move this issue to a more actively supported fork of sicktoolbox.
Hello,
We have been using the swri sicktoolbox and sicktoolbox_wrapper drivers for the NAV350 sensor for about two years now to retrieve pose data from the sensor. Recently an application has called for using the reflector and raw scan information from the sensor. However, this data appears to be available only in particular regions of the map.
If I move the sensor outside of a few regions of the map, the reflector transforms are no longer published and the scan data, while still being published, goes stale (identical msgs). If I start the driver in one of these "dark" areas (the majority of the map), I receive the following empty
LaserScanmsgs.I suspect that this may be an issue with the driver, because I did not notice any significant change in the packet sizes (via Wireshark) between the areas where the ROS driver publishes all the information (pose, reflector TFs, and scan data) and those in which it only publishes pose.
It was suggested to me that the scanner may not publish the information when there is a shortage of visible reflectors, however in the small region where all information is present (6" radius), we are seeing all 9 reflector frames. Upon leaving the region, all 9 abruptly disappear, as seen in the node DEBUG output below.
sicktoolbox: euler_integration sicktoolbox_wrapper: indigo-devel
Nodes running:
Running ROS Indigo