How to compare Path Planning Algorithms using Gazebo?

asked 2023-02-07 21:46:00 -0500

StarWarsFan17 gravatar image

Currently, I am trying to compare multiple path planning algorithms (Dijikstra's, A, A variants, etc.) using Gazebo as a way to verify the performance of different algorithms. In other words, although one can draw conclusions about path-planning algorithms through theoretical analysis, the study I am doing looks to draw a conclusion on which algorithm is optimal through application-based means. Hence, this is why I am using Gazebo to do so.

I was wondering what metrics I should look at to compare these algorithms, and how I would measure these metrics using any built-in tools Gazebo has. I'd like to get at least 3-4 metrics that I could measure and plot. I am more than okay with downloading any external add-on tools as well if that works best for what I am trying to do.

I plan on using standard maze-like environments built with Gazebo's world-building tools, and I also plan on using a differential drive unmanned ground vehicle for the study. I am using ros2 and my path planning program runs before the UGV actually moves (I am not using SLAM, using a OpenCV camera in a static environment). With that in mind, would it be better to test which path planning algorithm works best while a SLAM program is running? And if so, how would I do that? With that being said, lease let me know if I should provide any other details. Thank you very much, and I look forward to hearing feedback on this!

edit retag flag offensive close merge delete