GX5 publishes incorrect data on EKF topic

[ilansmolySR]

Describe the bug
Problem in /ekf/imu/data topic with 3DM-GX5-AHRS:
The robot is not moving but the linear accelearion is far from 0.
The roll, pitch are fixed to 0, and the yaw is fixed to 90.

When we connect 3DM-GV7-AHRS everything seems normal

But we need compatability with both types of sensors.

To Reproduce

1. Launch Node with command roslaunch microstrain_inertial_driver microstrain.launch
2. See Error

Environment (please complete the following information):

• OS: Ubuntu
• ROS Version: noetic
• Sensor(s): 3DM-GX5-AHRS
• docker environment

Launch Parameters
If you have an updated params.yml file, please attach it to this issue.

[ilansmolySR]

params.txt

[robbiefish]

Hi @ilansmolySR,

The reason you are seeing different behavior between the sensor has to do with the fact that the heading for the two devices is initializing differently

In your params.txt, you have the following line:

# This line sets the heading source for the GX5-AHRS to None
filter_heading_source : 0

You also have this line:

# This line sets the heading source for the GV7-AHRS to Magnetometer
filter_auto_heading_alignment_selector : 4

Since you are looking at the Linear Accel data quantity which attempts to factor out gravity, the device needs to have a somewhat reasonable estimate for the device's orientation in order to properly factor out gravity.

In order to resolve this issue, I would suggest setting filter_heading_source back to 1