navsat_transform broadcasting strange map->utm tf

edit

Hello there !

I'm trying to fuse data from different sensors (IMU, GPS and a rotating lidar) into an extended Kalman filter provided by robot_localization in order to drive my rover using move_base. I'm running ros indigo on ubuntu 14.04.

With the right packages i was able to publish three topics for my three sensors :

• /gps/fix : sensor_msgs/NavSatFix
• /imu/data : sensor_msgs/Imu
• /pose_with_covariance_stamped : geometry_msgs/PoseWithCovarianceStamped (let's call it /pose)

Here's how my navsat_transform_node and ekf_localization node are set :

navsat_transform_node

• Input : /imu/data, /gps/fix, /odometry/filtered
• Output : /odometry/gps

ekf_gps

• Input : /imu/data, /pose, /odometry/gps
• Output : /odometry/filtered

According to REP105 and what I've read in this forum I think my frames are set correctly, see the tf tree below:

image: (in the link below if not working)

Now when i plot my topics in rviz i see /odometry/filtered drifting to infinity and beyond and i don't understand why ... Looking at my tf tree it looks like the map->utm transform is bugged, however i got this message in my navsat_transform terminal :

INFO[1470904382.578813798]: World frame->utm transform is Origin: (-4945736.2295890189707 677781.95941346196923 -470427.19763114472153) Rotation (RPY): (0.085877088202917686854, -0.066475908266541719471, -1.5762575291089457874)

My initial pose doesn't look right, as for my orientation i've offseted my imu and my robot is almost pointing north so that looks correct.

I have rosbag all my topics and put my launch files link:here or link:here , i'd be glad if you could have a look at it and give me a clue of what i'm doing wrong ! I made the rosbag while the robot wasn't moving and pointing almost north.

Thank you for your time.

EDIT 1

Correcting the coma to a dot in value="0,00436332" (in navsat.launch) didn't change the drifting