Ask Your Question

Revision history [back]

click to hide/show revision 1
initial version

There are a couple of problems that I see in your parameter set:

1) If you really do intend to avoid obstacles based only on localization and a static map, you'll need to make sure that the local costmap is set up to run with that information. Otherwise, the local planner won't know about those static obstacles. If the map is small, you'll probably just be able to have the local costmap have the same parameters as the global costmap. However, if the map is going to be large, you'll probably want to feed the local costmap tiles of the global map to make sure computing the cost function for the map doesn't get too expensive.

2) The velocity limits you've set are extremely limiting. It'll be hard for the robot to follow any global path when its rotational velocity is limited to 0.07 radians/second and its translational velocity is limited to 0.07 meters/second. To make those limits work, I'd expect you'll have to play around a lot with the trahectory scoring parameters. Personally, I've never run navigation with such tight constraints.

There are a couple of problems that I see in your parameter set:

1) If you really do intend to avoid obstacles based only on localization and a static map, you'll need to make sure that the local costmap is set up to run with that information. Otherwise, the local planner won't know about those static obstacles. If the map is small, you'll probably just be able to have the local costmap have the same parameters as the global costmap. However, if the map is going to be large, you'll probably want to feed the local costmap tiles of the global map to make sure computing the cost function for the map doesn't get too expensive.

2) The velocity limits you've set are extremely limiting. It'll be hard for the robot to follow any global path when its rotational velocity is limited to 0.07 radians/second and its translational velocity is limited to 0.07 meters/second. To make those limits work, I'd expect you'll have to play around a lot with the trahectory trajectory scoring parameters. Personally, I've never run navigation with such tight constraints.