ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange
Ask Your Question
0

costmap_2d::ObstacleLayer meets problem with calculating cost value of obtacles inside footprint

asked 2022-09-12 22:15:28 -0600

N.N.Huy gravatar image

Hello everyone, I am using costmap_2d::ObstacleLayer to calculate local costmap of 2D laserscan. Normally, the obstacles inside footprint are ignored and not be calculated cost value. However, sometimes, whether the robot moves or stands still, somehow, the costmap_2d::ObstacleLayer still calucates cost value of these obstacles, especially those near the footprint boundary. You can see my photo photo link. I have tried spatio_temporal_voxel_layer, and I met the same problem. I guess there are some missing in logic of implementation of these plugins.

Anyone here solved the problem ? Thanks in advance and have a nice day!

edit retag flag offensive close merge delete

1 Answer

Sort by ยป oldest newest most voted
0

answered 2022-09-13 07:38:29 -0600

Mike Scheutzow gravatar image

As a workaround, some people set the config parameter <obstacle layer>.footprint_clearing_enabled to true.

The idea is that if your robot already occupies that space, then there wasn't really an obstacle there. Just be careful how this feature interacts with robot footprint padding.

edit flag offensive delete link more

Comments

Do you know the reason of this ?

N.N.Huy gravatar image N.N.Huy  ( 2022-09-13 23:22:02 -0600 )edit

You are asking why there are false objects left behind on the obstacle layer? No, I don't know for sure. I suspect it is because the base_link update is out of sync with the lidar update.

Mike Scheutzow gravatar image Mike Scheutzow  ( 2022-09-14 07:15:01 -0600 )edit

As in my Photo, the laser dot is still inside the footprint while there is an costmap dot near the laser dot, but outside the footprint. When the robot moves, the costmap dot appears frequently beside the laser dot. This situation happens both in real robot and gazebo simulation. Anyway, thanks for your idea ! If you have any more evidence, please let me know.

N.N.Huy gravatar image N.N.Huy  ( 2022-09-17 04:13:11 -0600 )edit

Question Tools

1 follower

Stats

Asked: 2022-09-12 22:15:28 -0600

Seen: 94 times

Last updated: Sep 13 '22