Ask Your Question

Obstacle Avoidance vs Detection in Navigation Stack

asked 2014-11-14 16:48:27 -0500

ROSCMBOT gravatar image

updated 2014-11-14 16:49:53 -0500


I was wondering if any sort of obstacle detection is happening in the navigation stack. As I understand, when an obstacle is appearing within laser range, it's updated in the local/global costmap, and the robot is doing obstacle avoidance by keep generating a global path (A* or Dijkstra) and moving along the new path, but does the robot know whether a new obstacle appeared on its path or not (or Does the robot detect the Obstacle?)


edit retag flag offensive close merge delete

1 Answer

Sort by ยป oldest newest most voted

answered 2014-11-14 20:09:04 -0500

fergs gravatar image

The robot doesn't really "know" or "reason" much about obstacles. It simply knows that "something" is at a particular location in the costmap and therefore it cannot go through it, but the default costmaps only reason about the occupancy of a particular grid, it doesn't do anything at the object level.

That said, there has been some work on higher level stuff. David Lu originally started on the layered costmap approach so that he could integrate a layer that understands and tracks where people are, allowing more reasoning about the obstacles. So, it is possible to do this kind of object tracking, but it's not part of the default navigation components.

edit flag offensive delete link more


Thanks a lot. If we don't really care about the exact location of the obstacles, but merely the fact that some new obstacles appeared on the cost map, can we find this out by keep comparing the old and new (global?) costmaps?

ROSCMBOT gravatar image ROSCMBOT  ( 2014-11-14 21:14:53 -0500 )edit

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Question Tools



Asked: 2014-11-14 16:48:27 -0500

Seen: 902 times

Last updated: Nov 14 '14