Hi Guys,

The setup is like this, I have a MobileRobots PeopleBot. I wrote a
playerstage wrapper for ROS so that ROS running on my laptop can communicate
with the player server on the robot. The connection is through a wired
ethernet.

I am seeing a strange case of tf failure here. The transform from base_link
to base_laser fails unless I subtract 1.75 seconds from the timestamp of
laser transforms!

I can see that the messages from odometery and laser scan are coming at
almost the same rate. Even the timestamps are almost identical and I am
timestamping these messages on the laptop.

The workaround of subtracting 1.75 seconds from the laser timestamp fixes
the base_link to base_laser transform but now when I try to run the
2dnavigation stack, the transform from map to base_laser won't happen
because the messages are too old. Even increasing the tolerance to something
like 2 seconds doesn't help as the buffer gets filled with old messages!

I guess my hack of subtracting 1.75 seconds at the first place was a wrong
move and I want to know if anybody had faced any such problem and if yes
then how did you debug it?

I have tried all the debugging techniques mentioned in the tf tutotrial on
ROS's website. Everything looks fine in view_frames or roswtf.

One more piece of information which might be helpful is that I am running
Ubuntu on VmWare on MacBook Pro. I don't know if virtual machine has
something to do in this, may be timing issue!

If you need any more info to help me out, please let me know and I will
furnish the same.
-- 
View this message in context: http://ros-users.122217.n3.nabble.com/Tf-failure-while-working-with-a-Pioneer-P3DX-tp2003281p2003281.html
Sent from the ROS-Users mailing list archive at Nabble.com.