IMU Linear Accelerations do not work correctly

[s1887252]

As stated at this link (in a different context but about the same topic)

https://answers.ros.org/question/12616/interpretation-of-imu-data-from-the-gazebo_ros_imu-plugin/

the UUV IMU implementation (as far as the Linear Accelerations are concerned) should be not correct.

Is it really like that?

[musamarcusso]

Hi @s1887252, the IMU is adapted from the Rotors Simulator and basically reads the linear acceleration from link https://github.com/uuvsimulator/uuv_simulator/blob/master/uuv_sensor_plugins/uuv_sensor_ros_plugins/src/IMUROSPlugin.cc#L183 But I recall having problem with the linear acceleration from Gazebo using the ODE engine with the quick algorithm. I can try changing the default algorithm to world to see if it behaves differently.

[HashirZahir]

I am not sure if this is related, but gazebo recently updated its plugin to report IMU orientation from world frame. I was facing some issues in my other project due to its old behaviour in terms of body orientation vs world orientation, but this pull request claims to fix it.

https://github.com/ros-simulation/gazebo_ros_pkgs/pull/1051