You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
While reviewing the code, I noticed that commands are generated alongside GPS data during the loading process. However, in CARLA, the presence of other interfering vehicles may affect the agent's path. If the agent continues to follow the previously generated commands despite interference, this could result in the "Agent deviated from the route" error.
Since model test is seriously affected by this problem. Do you have any plans to address this issue? Looking forward to your replay.
The text was updated successfully, but these errors were encountered:
Hi, this is a different concept. Current behavior is a real-time decision. Autonomous driving agents require a route or command(which derived from route) to tell them where to go.
One question I have is, will I update my cmd right away after I go over the near_node? I had predicted that I would turn right before I started to turn straight. I did not know whether the near_node was passed and caused the CMD to be updated. However, the vehicle did not travel along the correct route.
While reviewing the code, I noticed that commands are generated alongside GPS data during the loading process. However, in CARLA, the presence of other interfering vehicles may affect the agent's path. If the agent continues to follow the previously generated commands despite interference, this could result in the "Agent deviated from the route" error.
Since model test is seriously affected by this problem. Do you have any plans to address this issue? Looking forward to your replay.
The text was updated successfully, but these errors were encountered: