renweiros's profile - activity

Due to the comment length limit, I have to split my response into two comments. To resolve or at least alleviate the iss

Thank you for the pointer. Actually, I have tried the Husky frontier exploration demo given at http://wiki.ros.org/husky

This also works under kinetic! Thank bhaskara for the answer:)

I found one possible cause for this issue. The footprint config in navigation_stage/move_base_config/costmap_common_params.yaml doesn't match the sample config in frontier_exploration/launch/global_map.launch.

The footprint config in costmap_common_params.yaml:

#The footprint of the robot and associated padding
footprint: [[-0.325, -0.325], [-0.325, 0.325], [0.325, 0.325], [0.46, 0.0], [0.325, -0.325]]
footprint_padding: 0.01

The footprint config in global_map.launch:

footprint: [[0.1, 0.0], [0.0, 0.1], [-0.1, 0.0], [0.0, -0.1]]
robot_radius: 0.10

I tried changing the one in global_map.launch to match the one in costmap_common_params.yaml and started getting better results, i.e., the exploration won't fail quickly. But the exploration will often fail because the robot can't move, especially when it tries exploring a relatively large unknown area. Does anyone have any suggestion? Many thanks in advance.

I am hitting the same issue when I run the frontier_exploration demo given at the wiki page:

http://wiki.ros.org/frontier_explorat...

I followed the instructions given at the frontier_exploration wiki page:

http://wiki.ros.org/frontier_explorat...

roslaunch navigation_stage move_base_gmapping_5cm.launch
roslaunch navigation_stage move_base.xml
roslaunch frontier_exploration global_map.launch

After I specified an exploration region in rviz, the robot will start exploring but soon it will stops.

From the rviz and stage, I observed that the robot most of the times got stuck to some obstacle and failed to recover, but sometimes it was in a open space, not close to any obstacle, and still couldn't move.

In the frontier_exploration terminal, I always saw an error message:

[ERROR] [1491936923.242211778, 124.600000000]: Failed to move

In the move_base terminal, I saw a related error message when the robot got stuck:

[ERROR] [1491936126.596561945, 221.100000000]: Aborting because a valid plan could not be found. Even after executing all recovery behaviors

Has anyone else encounter a similar issue? Do I need to change some configurations/parameters of move_base and exploration_frontier? Any suggestion or help will be highly appreciated.

Run the demo Explore the Gazebo world. Check the tf tree via rqt_tf_tree and obtain the output as below:

rqt_tf_tree output

You can see that the tf of base_footprint -> base_link is published by the node robot_state_publisher.

But "tf_monitor" shows "unknown_publisher" for base_footprint -> base_link.

RESULTS: for base_footprint to base_link
Chain is: base_link -> base_footprint
Net delay     avg = 1114.58: max = 1164.52

Frames:
Frame: base_footprint published by unknown_publisher Average Delay: 0.00476341 Max Delay: 0.02

All Broadcasters:
Node: unknown_publisher 124.051 Hz, Average Delay: 0.00459184 Max Delay: 0.02

Also in the tf tree drawn by "rosrun tf view_frames", there is no tf from base_footprint to base_link.

tf view_frames output

Also notice that compared with the rqt_tf_tree output, there are many other tfs missing in the view_frames output.

Is this a bug in tf_monitor and view_frames? Or some configurations in the demo are incorrect?