Noisy gyroscope
| Required Info | |
|---|---|
| Camera Model | D435i |
| Firmware Version | 05.16.00.01 |
| Operating System & Version | Ubuntu 22.04 |
| Kernel Version (Linux Only) | 6.5.0-35-generic |
| Platform | PC |
| SDK Version | SDK2.0 (v2.50.0) |
Issue Description
My gyroscope data seems very noisy, as shown in the video. Is there any solution to fix this?
https://github.com/IntelRealSense/librealsense/assets/122732439/8abdabfa-346c-4e9c-9c94-be2286141dd8
Hi @B-Paweekorn RealSense SDK version 2.55.1 added an option to set the sensitivity of the gyro with the programming instruction RS2_OPTION_GYRO_SENSITIVITY. However, the sensitivity is not currently adjustable in the RealSense Viewer tool.
During tests in the Viewer that I performed, it seemed to be the up-down tilt action that contributed most to the gyro axes spinning very fast. Left-right turn of the camera produced more stable axis movement.
If using SDK 2.55.1 is not an option for you, it is also possible to alter the sensitivity of the gyro in C++ code by configuring an alpha value. This is demonstrated in the SDK's rs-motion C++ example.
https://github.com/IntelRealSense/librealsense/blob/master/examples/motion/rs-motion.cpp#L117-L118
Hi @B-Paweekorn Do you require further assistance with this case, please? Thanks!
Yes, please. If there are any additional solutions or if you could help identify the root cause of the problem, I would appreciate it.
| librealsense | 2.53.1 RELEASE |
| OS | Linux |
| Name | Intel RealSense D435I |
| Serial Number | 943222073465 |
| Firmware Version | 05.16.00.01 |
| Advanced Mode | YES |
| Camera Locked | YES |
| Usb Type Descriptor | 2.1 |
| Product Line | D400 |
| Asic Serial Number | 934123056297 |
| Firmware Update Id | 934123056297 |
I compared the movement of the gyro axes in the Viewer to the movement represented in the rs-motion C++IMU example program. The movement was significantly more stable in rs-motion.
The C++ code of rs-motion includes an 'alpha' value that can be configured to weight the IMU data towards the gyro (at the risk of drift) or towards the accelerometer (at the risk of higher sensitivity to disturbances). So it may be that rs-motion's apparent high stability means that the issue with the strongly fluctuating axes is in the RealSense Viewer specifically.
Are you able to test rs-motion to see how it performs with your camera, please?
https://github.com/IntelRealSense/librealsense/tree/master/examples/motion
I also encountered this problem. When installing version 2.50 in Ubuntu 20.04, IMU will have a lot of noise, but when installing version 2.51.1 in Windows 10, the noise is very small
Hi @ljysimple If you are experiencing an over-sensitive gyro then the gyro sensitivity feature added in librealsense 2.55.1 may be helpful if it is possible for you to update your SDK version.
Issues in SDK versions older than the current latest one that require code changes to be solved cannot be retroactively fixed by Intel though, unfortunately.
A key difference betwen the Linux and Windows versions of librealsense is that the Linux version runs on a UVC backend whilst the Windows version uses a different backend called Microsoft Media Foundation.
Thank you. The 2.50 version I downloaded mainly corresponds to the 2.32-ros version. Is there any corresponding realsense-ros version for the 2.55 version?
2.51.1 can be used with wrapper 2.3.2 as it is very similar to 2.50.0. Development on the RealSense ROS1 wrapper ceased after 2.50.0 and it has not been updated for more recent librealsense versions.
If you install librealsense 2.51.1 on Ubuntu then the 2.3.2 wrapper will need to built again afterwards.
Development of the RealSense ROS2 wrapper continues and there is a ROS2 wrapper for 2.55.1.
My version is ros1. After 2.51 is installed, it can support 2.3.2?
Yes, 2.51.1 is compatible with 2.3.2 as 2.51.1 only had minor changes from the previous version and so remains compatible with 2.3.2.
Thank you for your answer, I will try to solve it as soon as possible
Hi @B-Paweekorn and @ljysimple Do either of you have an update about this case that you can provide, please? Thanks!
Case closed due to no further comments received.