ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange
Ask Your Question

Revision history [back]

click to hide/show revision 1
initial version

Obstacle avoidance for variable obstacles is done by the local planner in the local costmap scope. What isnt mapped isnt taken into consideration by the global costmap and planner. It is also not ment to replan as there is no change expected.

I dont see where you actually upload the costap yaml files to the parameter server in the launchfiles. Also starting the map server before move_base might be better (change position in launchfile). A rosparam dump might be usefull. With all the yaml files and namespaces used to launch the files, quickly something can get messed up. And it seems everybody does it differently. I would expect move_base.launch to be the problem (usually its there where the yamls get uploaded and planners chosen). The discribed behaviour suggest a local problem with either costmap or planner. Most likely the obstacle_layer not marking.

Obstacle avoidance for variable obstacles is done by the local planner in the local costmap scope. What isnt mapped isnt taken into consideration by the global costmap and planner. It is also not ment to replan as there is no change expected.expected. You can think of the concept analog to a car with navigation system. The navigation system is the global_planner, plans once, and gives rough directions to the driver. The driver is analog to the local_planner who takes care of not crashing into other cars, stay in the lane, etc.

I dont see where you actually upload the costap yaml files to the parameter server in the launchfiles. Also starting the map server before move_base might be better (change position in launchfile). A rosparam dump might be usefull. With all the yaml files and namespaces used to launch the files, quickly something can get messed up. And it seems everybody does it differently. I would expect move_base.launch to be the problem (usually its there where the yamls get uploaded and planners chosen). The discribed behaviour suggest a local problem with either costmap or planner. Most likely the obstacle_layer not marking. marking.

Obstacle avoidance for variable obstacles is done by the local planner in the local costmap scope. What isnt mapped isnt taken into consideration by the global costmap and planner. It is also not ment to replan as there is no change expected. You can think of the concept analog to a car with navigation system. The navigation system is the global_planner, plans once, and gives rough directions to the driver. The driver is analog to the local_planner who takes care of not crashing into other cars, stay in the lane, etc.

I dont see where you actually upload the costap yaml files to the parameter server in the launchfiles. Also starting the map server before move_base might be better (change position in launchfile). A rosparam dump might be usefull. With all the yaml files and namespaces used to launch the files, quickly something can get messed up. And it seems everybody does it differently. I would expect move_base.launch to be the problem (usually its there where the yamls get uploaded and planners chosen). The discribed behaviour suggest a local problem with either costmap or planner. Most likely the obstacle_layer not marking.marking. Try visualizing the local_costmap and pathplans in rviz. This should give a quick understanding of whats going wrong.