Pose and Twist Covariances in the topic /realsense/odom

[SMish7]

Hi,

I intend to use the output of the RealSense SLAM node as an input to a Unscented Kalman Filter for doing some sensor fusion. For that, I need the pose and the twist covariance matrices, but in the topic named /realsense/odom, all the elements in the covariance matrices are set to zero. Is there any way to access the covariance matrices which result from the SLAM algorithm implement on the Intel Euclid Realsense device.

Thanks.

[idata]

Hello SubMarIn,

Thank you for your interest in the Intel® RealSense™ Technology.I'll be needing more to time to come up with an answer to your question. As soon as I have relevant information I'll post it here.Thank you for your patience.Regards,Andres V.

[SMish7]

Hey Andres,

Thanks for the response. I am eagerly waiting for an answer!

Regards,

S.

[SMish7]

Hi Andres,

Since you have implemented the slam algorithm, you will be definitely using the covariance matrix of the system states. A sensor no matter how accurate has some noise and uncertainty and knowing this uncertainty will help me in using the euclid sensor effectively in some data fusion algorithm. I will be glad if you let me know how to access the covariance matrix.

Regards,

S.

[idata]

Hello SubMarIn,

Thank you for providing a deeper explanation of why you want get access to the covariance matrix.We are still investigating, as soon as I have the corresponding information I'll contact you through this post.Thank you for your patience.Have a nice day.Regards,Andres V.

[SMish7]

Thanks Andres!

[idata]

Hello SubMarIn,

After further investigating, I have to tell you that the covariance matrices are not exposed in the SLAM library API, which means that they can't be accessed.Have a nice day.Regards,Andres V.

[SMish7]

Thanks for the information!!