In many of the ZAM_Urban Scenarios, a solution is found using gfbs / astar. The driveabitly checker then returns an invalid result.
this is the case:
https://commonroad.in.tum.de/submissions/ranking/KS2:SM1:ZAM_Urban-2_1:2020a
How to reproduce:
Run e.g. the commonroad search repo with gfbs algorithm on the scenario.
One way to solve the scenario: Expand the “standing still” nodes first or use a motion primitive with zero speed only. (solves around 400 out of 2077 2020a Scenarios alone)
Maybe there is an issue with the goal definition here:
- ZAM_Urban-2_1
- ZAM_Urban-3_1