How does the DWA planner handle continuously updated goals?

asked 2018-07-25 11:07:01 -0500

xjman530
1 ●2 ●3 ●4

updated 2018-07-25 11:42:06 -0500

Hello! I would appreciate some assistance since I am pretty new to ROS. I am running ROS indigo with ubuntu 14.04.

I am continuously updating the goal position of my robot (arlo-bot). I am wondering how the DWA_Local_Planner, and even the Global Planner, handles trajectory (path) formulation if I continuously publish the goal before the robot reaches the goal (move_base_simple/goal topic). Does it simply ignore the previous goal and recomputes the optimal trajectory for the newest goal? Or does it query all the goals until it reaches each goal, acting like a planning path? I am trying to determine how to set up my code so the robot can make the smoothest and fastest transitions between goals so that it does not stop to re-calculate the trajectory for every updated goal.

Thank you.

edit retag flag offensive close merge delete

Comments

On the other hand, I'm also worried the robot will not move if you continuously send it an updated path plan. Is there away to just append a via-point to the current plan instead of sending the planner the entire path_plan over and over again? I'm afraid this is why the robot is not moving. Thanks.

xjman530 ( 2018-08-02 15:33:09 -0500 )edit

add a comment

0

answered 2018-07-25 19:24:03 -0500

stevejp

929 ●1 ●8 ●18

Does it simply ignore the previous goal and recomputes the optimal trajectory for the newest goal?

This is what will happen if you continuously send goals. The move_base simple action server will receive the new goal, send it to the global planner to generate your global plan, and then the global plan is passed off to the local planner for path following / obstacle avoidance etc. If you're looking for more flexibility in your navigation strategy, you should check out the move_base_flex package.