Feature Request: Flag for changing NED to ENU frame

[austin-InDro]

The current IMU uses the NED frame.

I believe it would be a nice feature if the IMU could be automatically set up to be in the ENU frame by doing the following to the acceleration, gyroscope, and magnetometer data as the ROS navigation stack requires IMUs to be in the ENU Frame,.

    // set linear acceleration
    msg->linear_acceleration.x = last_accel_x_;
    msg->linear_acceleration.y = last_accel_y_;
    msg->linear_acceleration.z = last_accel_z_;

    // set angular velocities
    msg->angular_velocity.x = last_gyro_x_;
    msg->angular_velocity.y = last_gyro_y_;
    msg->angular_velocity.z = last_gyro_z_;

    imu_pub_->publish(std::move(msg));

    // Fill out and publish magnetic message
    mag_msg->header.frame_id = frame_id_;

    mag_msg->header.stamp = ros_time;

    mag_msg->magnetic_field.x = last_mag_x_;
    mag_msg->magnetic_field.y = last_mag_y_;
    mag_msg->magnetic_field.z = last_mag_z_;

to

    // set linear acceleration
    msg->linear_acceleration.x = last_accel_y_; 
    msg->linear_acceleration.y = last_accel_x_;
    msg->linear_acceleration.z = - last_accel_z_;

    // set angular velocities
    msg->angular_velocity.x = last_gyro_y_;
    msg->angular_velocity.y = last_gyro_x_;
    msg->angular_velocity.z = - last_gyro_z_;

    imu_pub_->publish(std::move(msg));

    // Fill out and publish magnetic message
    mag_msg->header.frame_id = frame_id_;

    mag_msg->header.stamp = ros_time;

    mag_msg->magnetic_field.x = last_mag_y_;
    mag_msg->magnetic_field.y = last_mag_x_;
    mag_msg->magnetic_field.z = - last_mag_z_;

[clalancette]

I agree that this would be a good feature to have. We added it to some other drivers in the past; see KumarRobotics/imu_vn_100@4441de2 for instance.

I'll warn that I don't think the above math is correct for converting NED to ENU. As far as I remember (it's been a while), you need to do something like https://github.com/KumarRobotics/imu_vn_100/blob/3ada41525b83b955f3d5a229c3d1f67dc31e7200/src/imu_vn_100.cpp#L722-L724 .

[austin-InDro]

Thanks for the link @clalancette, I don't believe that's needed for the Phidget_spatial library though (sorry, I forgot to specifiy which library I was talking about). There is not quaternion published, I use the imu_filter_madgwick fo that. I've found a few links mentioning that this way works for the conversion. Here is one linked,
mavlink/mavros#49

However, the link posted I believe is correct once the quaternion has been generated from the acc, gyro and mag values.

[mintar]

Closing this as resolved. As you stated, the phidgets spatial does not publish orientations (yet), so NED / ENU / ... doesn't apply here. Feel free to reopen if necessary!