Clarification on CHOMP Behavior as an Adapter After OMPL

[angeelalg]

I have successfully integrated CHOMP as a post-processing optimizer following OMPL in my MoveIt2 setup for a UR10e arm. While I have managed to get it operational, I'm encountering behavior that I don't fully understand and would appreciate some insights.

Specifically, CHOMP consistently terminates after the first iteration during my tests, irrespective of the scenario. Additionally, I keep seeing messages similar to "cCost 59.884768 over threshold 0.070000", suggesting some form of cost thresholding is involved in the termination condition, but the documentation hasn't made this clear to me.

Given that the termination seems to be cost-related, I attempted to adjust parameters such as ridge_factor to influence the optimization process. Increasing ridge_factor does indeed lead to more iterations, but CHOMP then fails to find a collision-free solution, perpetually iterating without success.

This has led me to several questions:

• What exactly determines CHOMP's termination condition, especially when used as an adapter after OMPL? Is it purely cost-based, and if so, how is this cost calculated?

• The role of ridge_factor and other parameters in influencing CHOMP's behavior and termination. How can these be effectively optimized to ensure a balance between iteration count and finding feasible, collision-free solutions?

I've followed the documentation and experimented with different configurations, but the specific mechanics of CHOMP's termination and parameter optimization remain unclear. Any guidance or additional documentation on this matter would be highly valuable.

Thank you for your time and assistance.

[github-actions[bot]]

This issue is being labeled as stale because it has been open 45 days with no activity. It will be automatically closed after another 45 days without follow-ups.

[github-actions[bot]]

This issue was closed because it has been stalled for 45 days with no activity.