fazildgr8 / ros_autonomous_slam

ROS package which uses the Navigation Stack to autonomously explore an unknown environment with help of GMAPPING and constructs a map of the explored environment. Finally, a path planning algorithm from the Navigation stack is used in the newly generated map to reach the goal. The Gazebo simulator is used for the simulation of the Turtlebot3 Waffle Pi robot. Various algorithms have been integrated for Autonomously exploring the region and constructing the map with help of the 360-degree Lidar sensor. Different environments can be swapped within launch files to generate a map of the environment.
http://mohamedfazil.com
MIT License
203 stars 69 forks source link

wall avoidance and multi turtlebot support #5

Open hikashi opened 3 years ago

hikashi commented 3 years ago

Hi Fazil,

I was wondering is it possible to recreate multiple turtlebot3 to run the RRT exploration as shown in the original RRT exploration package?

Also, I managed to get the program running but somehow the turtlebot3 have the tendency to stuck at wall and often requires manual control to unfreeze the robot? Is there any fine tuning parameters provided? thanks.

Regards, Billy

PabloSerrano-jpg commented 7 months ago

Hello! It seems I am having the same issue with wall avoidance. Was any solution found? Thanks!

hikashi commented 7 months ago

Hi, I think you have to aggressively inflates the costmap to solve this issue. However, note that it might reduce the traversability of narrow passage.

PabloSerrano-jpg commented 7 months ago

Hi, I think you have to aggressively inflates the costmap to solve this issue. However, note that it might reduce the traversability of narrow passage.

I see... In our case we're a team developing a rover for a contest and part of the challenge is an autonomous exploration to find the shortest path. The testing area would be very open, so narrow passages aren't an issue. As for the costmaps, which cpp file would I need to edit in this case if its either the global or local rrt detector. Thank you!