[X] This issue is not a duplicate. Before opening a new issue, please search existing issues.
[X] This issue is not a question, feature request, or anything other than a bug report directly related to this project.
Description
We have noticed that the front Zed 2i in our robot has a substantial drift compared to the Zed 2i mounted in the rear of the robot. The robot is a fairly large (larger than a Clearpath husky) tank drive robot with the Zed 2is mounted on aluminum. The zed 2is are mounted approximately 330mm from the floor, both zeds are mounted facing directly forward (both zeds are mounted parallel to the ground plane but in opposite directions)
We noticed that while the robot is completely still there is a large drift (often above 15cm) only in the front zeds position estimate. There is also noticeable drift between the zeds while the robot is moving.
We have tried multiple Zed 2is mounted on the robot and mounted outside the robot. When the Zed 2i is mounted outside the robot it does not have these drift issues, no matter how close we move it to the robot.
We have tried
2 different cables in different ports
2 different zeds 2is
isolating any vibrations by turning off the lidar and adding damping material to the zed 2i mounting (this lidar is mounted fairly near the front zed)
turning the robot around so the rear zed is facing the environment the front zed was facing and vice versa
You can see below that one zed has a sub 3cm drift in it's x,y direction (called stereo_nx i.e. the rear camera) while the other camera (called stereo px i.e. the forward facing camera) has about a -20cm drift in y and a 5cm drift in x
I have collected a rosbag of the IMU data from both cameras on the robot, and the drift in this bag. In the bag we see that the y acceleration for the forward camera is spuriously much higher
but the x acceleration is about the same
The drift is slightly better in this bag (posted below) but still there is a lot of y drift
I can send this bag to someone but I don't know the easiest way to do that. It has the odom output as well as the imu and magnetometer.
Below is the output of the ros zed sdk wrapper when the it is running (in docker)
Observe the odometry estimate from our front and rear zed 2is
Expected Result
We would expect that the rear and front Zed2is drift about the same (seems to be in general under 0.5cm). We would also expect the same drift when the zed 2i is mounted on the robot as when it's not, but we do not observe any unusual drift when it's not mounted.
Actual Result
An unusually high (significantly over 1cm) at rest drift rate only when this camera is mounted on the robot (no matter which camera or which cable).
Preliminary Checks
Description
We have noticed that the front Zed 2i in our robot has a substantial drift compared to the Zed 2i mounted in the rear of the robot. The robot is a fairly large (larger than a Clearpath husky) tank drive robot with the Zed 2is mounted on aluminum. The zed 2is are mounted approximately 330mm from the floor, both zeds are mounted facing directly forward (both zeds are mounted parallel to the ground plane but in opposite directions)
We noticed that while the robot is completely still there is a large drift (often above 15cm) only in the front zeds position estimate. There is also noticeable drift between the zeds while the robot is moving.
We have tried multiple Zed 2is mounted on the robot and mounted outside the robot. When the Zed 2i is mounted outside the robot it does not have these drift issues, no matter how close we move it to the robot.
We have tried
You can see below that one zed has a sub 3cm drift in it's x,y direction (called stereo_nx i.e. the rear camera) while the other camera (called stereo px i.e. the forward facing camera) has about a -20cm drift in y and a 5cm drift in x
I have collected a rosbag of the IMU data from both cameras on the robot, and the drift in this bag. In the bag we see that the y acceleration for the forward camera is spuriously much higher
but the x acceleration is about the same
The drift is slightly better in this bag (posted below) but still there is a lot of y drift
I can send this bag to someone but I don't know the easiest way to do that. It has the odom output as well as the imu and magnetometer.
Below is the output of the ros zed sdk wrapper when the it is running (in docker)
Steps to Reproduce
Expected Result
We would expect that the rear and front Zed2is drift about the same (seems to be in general under 0.5cm). We would also expect the same drift when the zed 2i is mounted on the robot as when it's not, but we do not observe any unusual drift when it's not mounted.
Actual Result
An unusually high (significantly over 1cm) at rest drift rate only when this camera is mounted on the robot (no matter which camera or which cable).
ZED Camera model
ZED2i
Environment
Anything else?
Thank you for taking the time to look at this, I know it may not be directly a ZED SDK issue.