Skip to content

Commit

Permalink
fix some minor point
Browse files Browse the repository at this point in the history
Signed-off-by: Zulfaqar Azmi <[email protected]>
  • Loading branch information
zulfaqar-azmi-t4 committed Nov 20, 2023
1 parent 41f0bec commit a599acc
Showing 1 changed file with 1 addition and 2 deletions.
3 changes: 1 addition & 2 deletions planning/behavior_path_planner/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -177,9 +177,8 @@ The purpose of the collision assessment function in the Behavior Path Planner is

1. During candidate path generation, to ensure that the generated candidate path is collision-free.
2. When the path is approved by the manager, and the ego vehicle is executing the current module. If the current situation is deemed unsafe, depending on each module's requirements, the planner will either cancel the execution or opt to execute another module.
This function operates under the assumption that accurate data on the position, velocity, and shape of both the ego vehicle (the autonomous vehicle) and any target objects are available. It also relies on the yaw angle of each point in the predicted paths of these objects, which is expected to point towards the next path point.

The safety check process involves several steps. Initially, it obtains the pose of the target object at a specific time, typically through interpolation of the predicted path. It then checks for any overlap between the ego vehicle and the target object at this time. If an overlap is detected, the path is deemed unsafe. The function also identifies which vehicle is in front by using the arc length along the given path.
The safety check process involves several steps. Initially, it obtains the pose of the target object at a specific time, typically through interpolation of the predicted path. It then checks for any overlap between the ego vehicle and the target object at this time. If an overlap is detected, the path is deemed unsafe. The function also identifies which vehicle is in front by using the arc length along the given path. The function operates under the assumption that accurate data on the position, velocity, and shape of both the ego vehicle (the autonomous vehicle) and any target objects are available. It also relies on the yaw angle of each point in the predicted paths of these objects, which is expected to point towards the next path point.

A critical part of the safety check is the calculation of the RSS (Responsibility-Sensitive Safety) distance-inspired algorithm. This algorithm considers factors such as reaction time, safety time margin, and the velocities and decelerations of both vehicles. Extended object polygons are created for both the ego and target vehicles. Notably, the rear object’s polygon is extended by the RSS distance longitudinally and by a lateral margin. The function finally checks for overlap between this extended rear object polygon and the front object polygon. Any overlap indicates a potential unsafe situation.

Expand Down

0 comments on commit a599acc

Please sign in to comment.