Ask Your Question

Revision history [back]

modify goal move base

Hi, I'm using the ros navigation stack for an exploration task. I use NavFN as global planner and TrajectoryPlannerRos as local planner (DWA = true).

What I would like to do is the following:

-I send a goal through actionlib::SimpleActionClient (move_base_msgs::MoveBaseAction). This goal is characterized by a specific position and orientation.

-While the robot is moving toward the goal I would like to modify the desired goal orientation.

I tried to send a new goal with the same position but different orientation, the problem is that the robot plans again a new path which is tipically different from the original one.

I don't understand how to achieve my desired behaviour, which should not be so complicated since the global planner only computes a list of positions (without orientation) and once the robot has reached the goal one, the local planner aligns it with respect to the issued desired orientation.

Thank for your help.

modify goal move base

Hi, I'm using the ros navigation stack for an exploration task. I use NavFN as global planner and TrajectoryPlannerRos as local planner (DWA = true).

What I would like to do is the following:

-I send a goal through actionlib::SimpleActionClient (move_base_msgs::MoveBaseAction). This goal is characterized by a specific position and orientation.

-While the robot is moving toward the goal I would like to modify the desired goal orientation.

I tried to send a new goal with the same position but different orientation, the problem is that the robot plans again a new path which is tipically different from the original one.

I don't understand how to achieve my desired behaviour, which should not be so complicated since the global planner only computes a list of positions (without orientation) and once the robot has reached the goal one, the local planner aligns it with respect to the issued desired orientation.

Thank for your help.

UPDATE:

I'm implementing an exploration algorithm. My idea is to send as goal a pose from which a big unexplored area could be seen. While moving toward the goal the robot could explore this area "by chance" so a typical approach would be to abort the goal and plan again for a new one. I think that sometimes could be more convenient to keep the same goal (because maybe we have almost reached it) and request a different orientation (from which an unexplored area could still be observed). The problem in planning again with a different orientation is that due to the changes in the map and the not deterministic behaviour of the planner, the new plan could have a different shape with respect to the old one and, in order to follow it, strange maneuvers could be necessary.

By the way the answer and further trials convinced myself that this is the only way of achieving this.