Closed Alessio-Parmeggiani closed 5 months ago
Hi @Alessio-Parmeggiani
The covariance values from the topic /zedxm/zed_node/odom are filled only on the diagonal and the values are all equal to 1.00000somethingE-10, the matrix also does not change at all during the movement of the camera. The same is true also for the /zedxm/zed_node/pose_with_covariance topic. Is this expected? In the previous positional tracking version (now GEN_1) the covariance matrix was completely filled and values appeared more correct,as they were all different and change based on the movement of the camera.
This is an expected behavior. The covariance with GEN_2 is not yet available.
The second strange behavior happens randomly during the movement of the camera, in fact the camera stops publishing values on the /zedxm/zed_node/odom topic. I think it could happen when the camera get lost but if it is the case, it would be nice to have this published somewhere: as a warning on the terminal, or published on the /zedxm/zed_node/odom/status topic. Again, is this behavior expected?
Can you please check the status with the Diagnostic? You can use rqt
and the "Runtime Monitor" plugin.
I expect it to be SEARCHING
@Myzhar Hi, thank you for the response.
I checked the diagnostics messages and it says UNAVAILABLE when it stops publishing messages, normally it displays OK.
That means indeed that the visual tracking lost the information: https://www.stereolabs.com/docs/api/group__PositionalTracking__group.html#gae481924a311113375bb3a99748d79e17
That means indeed that the visual tracking lost the information: https://www.stereolabs.com/docs/api/group__PositionalTracking__group.html#gae481924a311113375bb3a99748d79e17
Ok, thank you, everything is clear now
Preliminary Checks
Description
Hi, I was testing the ros wrapper with the new GEN_2 positional tracking with ZED SDK 4.1 and I noticed two strange behaviors:
The covariance values from the topic
/zedxm/zed_node/odom
are filled only on the diagonal and the values are all equal to 1.00000somethingE-10, the matrix also does not change at all during the movement of the camera. The same is true also for the/zedxm/zed_node/pose_with_covariance
topic. Is this expected? In the previous positional tracking version (now GEN_1) the covariance matrix was completely filled and values appeared more correct,as they were all different and change based on the movement of the camera.The second strange behavior happens randomly during the movement of the camera, in fact the camera stops publishing values on the
/zedxm/zed_node/odom
topic. I think it could happen when the camera get lost but if it is the case, it would be nice to have this published somewhere: as a warning on the terminal, or published on the/zedxm/zed_node/odom/status
topic. Again, is this behavior expected?Steps to Reproduce
roslaunch zed_wrapper zedxm.launch
rostopic echo -c /zedxm/zed_node/odom
Expected Result
I expect the odometry output to be more similar to the one produced by the GEN_1 algorithm, i.e. the covariance matrix is all filled and with values that change, and the odometry should not stop randomly.
Actual Result
The covariance values are filled only on the diagonal and with static values. The odom topic stops publishing messages without any warning or special status in the odom/status topic.
ZED Camera model
ZED
Environment
Anything else?
No response