Saved in:
Bibliographic Details
Main Authors: Pfaffmann, Donald, Klusch, Matthias, Steinmetz, Marcel
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2510.07210
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914309847646208
author Pfaffmann, Donald
Klusch, Matthias
Steinmetz, Marcel
author_facet Pfaffmann, Donald
Klusch, Matthias
Steinmetz, Marcel
contents We present a novel hybrid learning-assisted planning method, named HyPlan, for solving the collision-free navigation problem for self-driving cars in partially observable traffic environments. HyPlan combines methods for multi-agent behavior prediction, deep reinforcement learning with proximal policy optimization and approximated online POMDP planning with heuristic confidence-based vertical pruning to reduce its execution time without compromising safety of driving. Our experimental performance analysis on the CARLA-CTS2 benchmark of critical traffic scenarios with pedestrians revealed that HyPlan may navigate safer than selected relevant baselines and perform significantly faster than considered alternative online POMDP planners.
format Preprint
id arxiv_https___arxiv_org_abs_2510_07210
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle HyPlan: Hybrid Learning-Assisted Planning Under Uncertainty for Safe Autonomous Driving
Pfaffmann, Donald
Klusch, Matthias
Steinmetz, Marcel
Robotics
Artificial Intelligence
We present a novel hybrid learning-assisted planning method, named HyPlan, for solving the collision-free navigation problem for self-driving cars in partially observable traffic environments. HyPlan combines methods for multi-agent behavior prediction, deep reinforcement learning with proximal policy optimization and approximated online POMDP planning with heuristic confidence-based vertical pruning to reduce its execution time without compromising safety of driving. Our experimental performance analysis on the CARLA-CTS2 benchmark of critical traffic scenarios with pedestrians revealed that HyPlan may navigate safer than selected relevant baselines and perform significantly faster than considered alternative online POMDP planners.
title HyPlan: Hybrid Learning-Assisted Planning Under Uncertainty for Safe Autonomous Driving
topic Robotics
Artificial Intelligence
url https://arxiv.org/abs/2510.07210