Hi Bünyamin,

First, could you ask this on answers.ros.org so that it can be more easily indexed and searched for in the future?

Second, have you tried running your gmapping with the rosparam "use_sim_time" set to true and "rosbag play --clock"? It looks like you might just have troubles because the timestamps in the bagfile for tf data correspond to when you recorded the data, but post-processing in gmapping will only be buffer tf data that has timestamps that correspond to whenever you are running the gmapping node. This tutorial has some guidance on running gmapping on logged data: http://www.ros.org/wiki/slam_gmapping/Tutorials/MappingFromLoggedData

- Eric

2011/4/27 "Yilmaz Bünyamin" <yilmaz.buenyamin@rwth-aachen.de>
Hello,
i posted before some problems while i was trying to create a map.

-First, to use slam algorithm u need a topic with a laser_scan msg.
This was my first to do.
Because nxt_robot_sensor_car is not publishing a topic with laser_scan.
I tried to write a Node wich is subscribing the topic ultrasonic_sensor and then converting the data and publishing a topic named base_scan with the msg type laser_scan.
The node is also turning the robot 390 degree. You can change that if u want.

I attached this node if anyone wants to look in to that code.
But before u look into that code, i am warning you, it is not good implementated.
Even i dont know if its working corretly.
So i would make me happy, if i get some feedbacks :)

-Next i recorded base_scan and tf in a bag file.(base_scan is given by the node that i have programmed,the /tf topic is given by nxtros...)
-now if i play the bag file the slam_gmapping node is givin warnings like:

[ WARN] [1303911122.112710800]: TF_OLD_DATA ignoring data from the past for frame /accel_link at time 1.30391e+09 according to authority /play_1303911121895057400
Possible reasons are listed at

-So what is my problem now?
-what should i do next or fix next?

Greetings,
Bünyamin



_______________________________________________
ros-users mailing list
ros-users@code.ros.org
https://code.ros.org/mailman/listinfo/ros-users