Fix body outputs, proper integration of imu_short and remove dynamic tf-broadcaster for static transform

[gkueppers]

Introduction

We have identified a few minor issues with our Ekinox Micro. First of all, we noticed that the body outputs are dependent on use_enu, which, in our opinion, does not make sense (see issue #66 ).

Furthermore, problems arose during the update when using the new imu_short message, which we have resolved with the appropriate changes.

We are open to feedback on the implementation or possible misunderstandings on our part.

Summary

• Make body outputs (/sbg/ekf_vel_body, /sbg/ekf_rot_accel_body) independent of the use_enu flag (Closing Influence on use_enu on /sbg/ekf_vel_body and /sbg/ekf_rot_accel_body #66)
• Enable publishing of ROS standard odom and imu message if imu_data OR imu_short is available
• Use areStampsClose to match imu_short message with other required inputs for imu/odom instead of hard comparison whether timestamp is the same (this was not the case with our Ekinox Micro)
• don't use a (dynamic) transform broadcaster for the static-tf odom_init_frame_id_ --> odom_frame_id_

Possible Open Todo's

The following tasks can be tackled after and based on your initial feedback, if this pull request is conceivable to be merged:

• Update the README
  • /sbg/imu_short is missing in published topics
  • Align section on Body/Vehicle Frame
• Remove magic-number in areStampsClose
• Switch back to time_stamp instead of header.stamp to match imu_short message with other required inputs for imu/odom