Using cost map inflation on future occupation distributions

asked 2015-05-18 09:16:02 -0500

JonaSc gravatar image

I'm writing a planner which plans with respect to predicted movements of dynamic obstacles. My planner adheres to the nav_core::BaseGlobalPlanner class and is therefore able to access the costmap_2DROS object.

The future occupancy distributions are given by nav_msgs/OccupancyGrid messages. I would now like to inflate the occupancy data provided with respect to the occupancy probability.

Given a centre cell which is occupied with Probability P, I would like to update all cells within the circumscribed region to P * Inflated_cost (unless their current cost is higher), where Inflated_cost is the cost which would be assigned to the cell given that the centre cell is a lethal obstacle.

1) Is it possible to use a costmap_2d object to do this calculation for me? I thought about creating a new costmap object, providing it with the static map and the predicted occupation and let it inflate the obstacles for me.

2) In case the answer to 1) is no: How does the inflation_layer actually inflate the obstacles? Is there any documentation on this?

edit retag flag offensive close merge delete