Magnetometer data is missing the covariance

[mllofriu]

Please provide the following information:

• OS: Ubuntu 20.04 in Tegra
• ROS Distro: Foxy
• Built from source or installed: Installed, built jackal packages from source
• Package version: ee4e3f6

Expected behaviour
The messages in /imu/mag have non-zero covariaces.

Actual behaviour
The data has the Zero matrix as covariance.

To Reproduce
Provide the steps to reproduce:

1. run jackal bringup
2. ros2 topic echo /imu/mag

[tonybaltovski]

@mllofriu The magnetometer didn't have covariance data from ROS 1 when it was only a Vector3 message. Do you have a specific use case for the data? We can look for the covariance for the sensor and add it to the next release of the firmware.

[mllofriu]

Some lidar SLAM systems require a 9-degree imu, which could be composed of the 6-degree imu and the magnetometer.

To be honest, I'm not sure how sensitive to the covariance any of these solutions (if at all) are:

• use a (stateless?) madgwick filter to get a 9-degree estimate from both sensors, and use that as the 9-degree imu
• for the algorithms allowing to read from /imu/mag directly, how sensitive they are to the reported covariance

All in all, I think this one is lower priority, unless you believe it affects the result of the /imu/data estimate (madgwick).