Robotics StackExchange | Archived questions

Navigation Stack path planning

Dear all.

I have a quadratic footprint shape differential drive mobile robot.

I have been using the navigation stack successfully, however I found some issues.

1.- Difficulties to navigate in very narrow spaces, the robot frequently gets trap and the global planner cancels. 

2.- The planned path from the current robot  to the goal positions is not calculated in the middle of the empty space which is situated  either between the obstacles or the inflated obstacles. Instead, the path is calculated following the shape of the inflated obstacles, in turn this makes the robot get trap causing the dropping of the global planner to the goal.

3.- The robot does not follow precisely the calculated path, there is always a shifted error.

4.- When the robot gets trap, instead of rotating and calculating a new path, it just moves a bit forwards and backwards and does not calculate any new path. Then, the global plan drops. 

Just wondering if some one can give me some tips on how to tackle the previous issues.

In advance, thank you for your kindly help :)

Asked by acp on 2014-09-17 06:19:40 UTC

Comments

Answers

Check out this "ROS Navigation Tuning Guide" it was quite useful: https://arxiv.org/pdf/1706.09068.pdf

Asked by Combinacijus on 2021-02-26 13:16:31 UTC

Comments