DWB_local_planner - Distance between the robot and the obstacle

Question

Dear all,

I use the DWB_local_planner and the robot gets very close to the obstacle while avoiding it.

I'm interested to configure the DWB_local_planner in order that the robot ALWAYS keep a certain distance (for instance 50 cm) to any obstacles.

In the TEB_local_planner there is a parameter called min_obstacle_dist for this issue. This distance is independent from the local_costmap inflation_radius.

I was wondering if you know how to configure the DWB_local_planner that the robot keeps a specific distance from the obstacle. In the DWB_local_planner parameter, as far as I understood, a specific parameter has not been dedicated for this issue.

Thanks,

---

Originally posted by Mobile_robot on ROS Answers with karma: 264 on 2019-01-24

Post score: 0

---

Original comments

Comment by David Lu on 2019-01-24:
What do you mean an override? A change via DWB that doesn't affect the costmap?

Comment by Mobile_robot on 2019-01-24:
I meant that I prefer to have access to the distance between the robot and the obstacle independently from the costmap. For example if the inflation radius is equal to 0.1m I can configure it to have a distance equal to 0.3m. Do you think it is possible?

Comment by David Lu on 2019-01-24:
I'm still not certain what you mean. Are you asking if you can just increase the inflation_radius parameter?

Comment by Mobile_robot on 2019-01-25:
Thanks for your reply David. I updated my question and I hope now it is clear for you.

Answer 1

So if you want the robot to NEVER go within X meters of an obstacle, I would change the robot's footprint to have an inscribed radius equal to X. This will make some feasible plans not feasible anymore.

If you want the robot to strongly prefer not going within X meters, you can increase the scale parameter on the obstacle critic in DWB.

If you want to control the logic even further, I would create a custom DWB critic that extends the MapGridCritic and use it to calculate the distance to the obstacles, and evaluate trajectories based on their distance to the obstacle.

---

Originally posted by David Lu with karma: 10932 on 2019-01-25

This answer was ACCEPTED on the original site

Post score: 1

---

Original comments

Comment by Mobile_robot on 2019-01-25:
Thank you David!