added another callback in imu_ned_enu based on euler angle changes. T…

[mzhouURI]

I updated the imu_ned_enu node.
The previous one has some problem in roll and pitch.
so i created another callback function where

• The roll, pitch yaw is first obtained form the Qua. (in NED frame)
• then we apply roll pitch and yaw offset (or reverse sign) to get the correct orientation in ENU.
• convert rpy back to Qua (in ENU) and replace the value in imu msg
• publish to a new topic /imu/data
• added a param for frame_id. so the new imu data in enu can have a different frame_id

I tested with alpha_rise_auv https://github.com/GSO-soslab/alpha_rise_auv
The orientation of baselink and imu links are correct, and the rotation are correct compared to the stonefish UI.

[incebellipipo]

1. Instead of using pi as 3.14..... you should use M_PI for pi, M_PI_2 for pi/2

2. You don't need to have a separate frame id, a different topic name would suffice. Unless the described orientation is also not changing.

3. I used this Check frame conversions (NED->ENU ROS) mavlink/mavros#49 (comment) to convert NED to ENU. To reduce the number of Quaternion -> Euler -> Quaternion conversions. I can try to make sure that can apply first. But your method looks rather safe and should work.

[mzhouURI]

I changed the 3.14.... related things to M_PI and M_PI_2.
Fixed a small complain for the target frame_id

[mzhouURI]

Update:
I will revert my change.
and have my way of converting imu in a new node.
user could define roll, pitch, yaw offset or reverse in rosparam.

Once it is tested, i will update and request a pull request review