imu/data_raw: axis switch within driver. Why?

[PhilippSchmaelzle]

I am new to this driver and I plan to use the ROS1 Version of it.

With commit f29f11e the suppport of the RAWIMUX is provided! Super cool, thx for this.
But i do not see the reason why the X- and Y axis are switched and negated.

link to code

Expected behavior
As this is a raw signal of the IMU and the log provides this as such (without internal transformation to RBV or USER frame) I would expect the values to be published in the IMU frame (imprinted on the IMU housing)

Or a similar implementation as you did with the CORRIMU log where you provide oem7 as well as rep 105 conform output.

For me the remapping of x = -y and y = -xis not obvious but maybe I just do not see it 😄

I hope you can help me with this.

Thanks for this otherwise great driver! As we used our own implementation until now.

Best regards

[Timple]

The request for the imu/data_raw came from us I think. They asked if they should adhere to REP-145: #26 (comment)

We responded that that should be the case, but somehow it was implemented different 🙁

[PhilippSchmaelzle]

Thx @Timple for the hint!
To be consistent within the driver I think the RAWIMUX data frame should also be configurable via oem7_imu_reference_frame_. Otherwise users of the oem7 frame do not have matching imu/data_raw values.

[novatel-applications-engineering]

Hi @PhilippSchmaelzle @Timple thank you for reporting this issue. We are currently investigating and will provide an update when/if a fix is required. Thank you!

[novatel-applications-engineering]

Hi @PhilippSchmaelzle @Timple we have resolved this issue in ROS1 branch: 8ecee3e

We will port this change over to ROS2 once confirmed that it is operating as expected.