Hi!
When planning, I've noticed an issue with how the exploration graph is being sampled. It tends to expand preferentially near the walls, forming a "U" shape when approaching the edge of the scan.
From the visualization, it seems that the algorithm might be misclassifying certain areas as non-free space, even though the TSDF representation suggests that there is ample free space available.
The TSDF looks like this:
As you can see, there is plenty of free space there.
Could this be related to how the planner thresholds free vs. occupied space and, are there any parameters that could be tuned to improve sampling behaviour in open areas?
Thanks in advance!
Hi!
When planning, I've noticed an issue with how the exploration graph is being sampled. It tends to expand preferentially near the walls, forming a "U" shape when approaching the edge of the scan.
From the visualization, it seems that the algorithm might be misclassifying certain areas as non-free space, even though the TSDF representation suggests that there is ample free space available.
The TSDF looks like this:
As you can see, there is plenty of free space there.
Could this be related to how the planner thresholds free vs. occupied space and, are there any parameters that could be tuned to improve sampling behaviour in open areas?
Thanks in advance!