The behavior_path_planner can not plan the right path in some positions as the target

[shulanbushangshu]

Checklist

• [X] I've read the contribution guidelines.
• [X] I've searched other issues and no duplicate issues were found.
• [X] I'm convinced that this is not my fault but a bug.

Description

When I set the different positions of target in the lane,the behavior_path_planner can not plan the right path in some positions. case 1:

case 2:

case 3:

case 4:

Expected behavior

The behavior_path_planner shall plan the right path in all positions of the lanes.

Actual behavior

The behavior_path_planner can not plan the right path when I set some different positions as the target in the lanes.

Steps to reproduce

1.Start autoware planning simulator;

2. Set initial position
3. Set target in the different positions in the lane(far from the initial position)

Versions

1.OS: Ubuntun 20.04 2.ROS2: Galactic 3.Autoware: Autoware.universe:main

Possible causes

No response

Additional context

No response

[ismetatabay]

Hi, can you please share map (pcd and lanelet) files? The lanelets which highlighted as I have shown below picture show that the route has been successfully generated by the mission_planner, I didn't see any problem in this case. If you're worried about not seeing the green path in rviz, the path will move forward as the vehicle move. The parameter that is forward_path_length limits the path . You can also increase the generated path by changing this parameter.

https://github.com/autowarefoundation/autoware.universe/blob/29e8b435a4416e5db0eb6c7da0c4b0457e6c529b/launch/tier4_planning_launch/config/scenario_planning/lane_driving/behavior_planning/behavior_path_planner/behavior_path_planner.param.yaml#L4

[shulanbushangshu]

@ismetatabay As you say, the route has been successfully generated by the mission_planner.Yes,I get what you mean.The distance from the target is more than 300m.But the problem is that final path of the behavior_path_planner is short(I print the size of the path and it shows that there are just 8 points in the path)

[shulanbushangshu]

@ismetatabay I test the code and show the information about the path of behavior_path_planner as below:

case1:

case 2:

[ismetatabay]

@shulanbushangshu This is weird thing that it shouldn't happen. At this case, if forward_path_length is 300m and backward_path_length is 5m, the number of path points created should be approximately 305/2 ≅152. To see if this situation is related to distance goal distance , I modified the sample .osm file of autoware as in the image below.

I've tried many different scenarios but haven't run into a problem. As expected, if the distance to the goal was greater than 300 meters, approximately 152 path points were generated on bpp.

Is it possible for you to share the lanelet2 .osm file?

[stale[bot]]

This pull request has been automatically marked as stale because it has not had recent activity.