We chose not to publish the raw accelerometer, magnetometer, and gyroscope data into a sensor_msgs/Imu topic mainly because we currently have no way of computing orientation from the 3 data sources. I can include a publisher for a sensor_msgs/Imu topic but it would only have the angular_velocity and linear_acceleration data fields populated. Would that solution work for you?
Publishing of a sensor_msgs::Imu topic was not added to version 3.3.3 of the ROS driver. We still have no means of accurately generating a orientation from the raw IMU streams in the ROS driver. As mentioned above I could add a topic with just the linear_acceleration and angular_velocity fields populated, but I was under the impression that did not have very much utility over the current IMU publishing scheme.
thanks for the reply. For me it plays a big role to have it as a sensor_msgs::IMU topic because in that way I avoid including the RawImuData.h file in my file for using the multisense_ros::RawImuData topic in my callback function. This is a big big issue and make things very complicated from our system management point of view. Just to give you an idea of how much this affects us: the way to use the multisense_ros message right now could be possible only by replicating (!!!) the RawImuData.h code in ours. And I would like to avoid this. So I believe that even without the orientation it's OK to publish such a topic! Please let me know if/when you will be able to integrate such a topic! Thanks a lot!!!