smac_planner: fix goal checking in getAnalyticPath()#2159
smac_planner: fix goal checking in getAnalyticPath()#2159mkolodziejczyk-piap wants to merge 1 commit intoros-navigation:foxy-develfrom
Conversation
SteveMacenski
left a comment
There was a problem hiding this comment.
Just a styling change but LGTM. Can you explain in a few sentences what the specific issue you hit? I'm curious how we didn't also run into this in our testing.
| for (const auto & node_pose : possible_nodes) { | ||
| const auto & n = node_pose.first; | ||
| if (!n->wasVisited()) { | ||
| if (!n->wasVisited() and n->getIndex()!=_goal->getIndex()) { |
There was a problem hiding this comment.
| if (!n->wasVisited() and n->getIndex()!=_goal->getIndex()) { | |
| if (!n->wasVisited() && n->getIndex() != _goal->getIndex()) { |
Just to be consistent in styling
|
IMO the issue arises, when the vehicle is crossing the goal but the orientation is not the goal orientation and after crossing this goal ompl plans to correct the yaw with some back and forth manouvers. In my case, this loop happend almost all the time near the goal, probably because (1) my controller was not tuned enought to follow the path exactly (so the initial "smooth" global plan turned into some wild manouvers to correct the yaw) (2) my vehicle has large turning radius |
Ooooooh, yeah I did not test for that. Thanks! Please make the update / |
|
@mkolodziejczyk-piap just following up here. Just need the small styling update & |
|
@mkolodziejczyk-piap just following up This is essentially done, just needs superficial update. If I don't hear from you, I'll merge in my own patch sometime next week. |
|
Closing - superseded by linked PRs that meet the Nav2 coding styling |
Basic Info
Description of contribution in a few bullet points