Skip to content
This repository has been archived by the owner on Feb 8, 2023. It is now read-only.

Noisy Linear Acceleration #16

Closed
stackSmasher294 opened this issue Oct 31, 2017 · 3 comments
Closed

Noisy Linear Acceleration #16

stackSmasher294 opened this issue Oct 31, 2017 · 3 comments

Comments

@stackSmasher294
Copy link

Hi,

I enabled streaming of gravity-removed linear acceleration by adding BG5_LINEAR_ACCEL_BODY to Binary Group 5 in the 'Stream()' method and reading back 'data.linearAccelBody' in the 'PublishData' method. However on visualising the acceleration in Rviz and also on looking at the contents of the topics, it seemed that the values are quite noisy even when the sensor is standing still.

I have set the VPE to be enabled and imu_compensated to be true. Even the non-gravity compensated acceleration seemed to be noisy on rviz. I have not been able to use the acceleration measurements from this IMU in my sensor fusion algorithm because of such noisy data.

Am I missing out on enabling certain flags/settings?

@kartikmohta
Copy link
Collaborator

Can you plot the acceleration values and show the amount of noise?

@stackSmasher294
Copy link
Author

stackSmasher294 commented Nov 1, 2017

On keeping the sensor still, the peak-to-peak jitter in acceleration value is within 0.03-0.08 m/s^2. Our application has a theoretical maximum of around 0.50 m/s^2. Given the small value of peak acceleration, the idle jitter is quite higher in comparison.

Is this amount of jitter nominal for vn-100?

@kartikmohta
Copy link
Collaborator

Here is a plot of the data I get when the IMU is stationary:
accel_x

Seems like the noise is similar to what you described. I believe that this amount of noise is pretty standard.

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants