Added documentation for asymmetric_inflation_layer#913
Added documentation for asymmetric_inflation_layer#913mbloechli wants to merge 8 commits intoros-navigation:masterfrom
Conversation
Signed-off-by: mbloechli <mbloechlinger@duatic.com>
| :``<asymmetric inflation layer>``.neutral_threshold: | ||
|
|
||
| ====== ======= | ||
| Type Default | ||
| ------ ------- | ||
| double 2.0 | ||
| ====== ======= | ||
|
|
||
| Description | ||
| Maximum perpendicular distance (m) from the path centreline. | ||
|
|
||
| Obstacles farther than this distance are ignored. |
There was a problem hiding this comment.
I think work a discussion: Why is this? If we're applying the inflation over some incremental window, why wouldn't we be applying it for the full update window?
There was a problem hiding this comment.
Assuming, that we are chaining the asymmetric_inflation_layer after the standard inflation_layer:
By ignoring obstacles farther away than neutral_threshold, we can save computation resources. Cells far away from the global path, we can assume to be irrelevant for our controllers local planning. I usually set the neutral_threshold to be the same as inflation_radius. So that all lethal_cells, that won't contribute a shift to the Voronoi border, don't have to be inflated a second time and can keep the same inflation as the standard inflation_layer already did.
There was a problem hiding this comment.
Got it, I didn't realize in reading this that this was chained. I suppose that's a conversation we still need to discuss in the main PR. I'm really slammed right now and onsite with a sponsor so I haven't gotten to that yet. My apologies and this is not forgotten or lost! This is a priority for me but I have to do what I have to do onsite while I am here
Signed-off-by: mbloechli <mbloechlinger@duatic.com>
Signed-off-by: mbloechli <mbloechlinger@duatic.com>
Signed-off-by: mbloechli <mbloechlinger@duatic.com>
| :``<asymmetric inflation layer>``.neutral_threshold: | ||
|
|
||
| ====== ======= | ||
| Type Default | ||
| ------ ------- | ||
| double 2.0 | ||
| ====== ======= | ||
|
|
||
| Description | ||
| Maximum perpendicular distance (m) from the path centreline. | ||
|
|
||
| Obstacles farther than this distance are ignored. |
There was a problem hiding this comment.
Got it, I didn't realize in reading this that this was chained. I suppose that's a conversation we still need to discuss in the main PR. I'm really slammed right now and onsite with a sponsor so I haven't gotten to that yet. My apologies and this is not forgotten or lost! This is a priority for me but I have to do what I have to do onsite while I am here
…icture with correctly cropped one, moved the picture to top of page Signed-off-by: mbloechli <mbloechlinger@duatic.com>
…icture with correctly cropped one Signed-off-by: mbloechli <mbloechlinger@duatic.com>
…t up cost_scaling_factor into left and right side Signed-off-by: mbloechli <mbloechlinger@duatic.com>
2 participants
Basic Info
Description of contribution in a few bullet points