Informed Reinforcement Learning for Situation-Aware Traffic Rule Exceptions

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Bogdoll, Daniel, Qin, Jing, Nekolla, Moritz, Abouelazm, Ahmed, Joseph, Tim, Zöllner, J. Marius
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908347792359424
author Bogdoll, Daniel
Qin, Jing
Nekolla, Moritz
Abouelazm, Ahmed
Joseph, Tim
Zöllner, J. Marius
author_facet Bogdoll, Daniel
Qin, Jing
Nekolla, Moritz
Abouelazm, Ahmed
Joseph, Tim
Zöllner, J. Marius
contents Reinforcement Learning is a highly active research field with promising advancements. In the field of autonomous driving, however, often very simple scenarios are being examined. Common approaches use non-interpretable control commands as the action space and unstructured reward designs which lack structure. In this work, we introduce Informed Reinforcement Learning, where a structured rulebook is integrated as a knowledge source. We learn trajectories and asses them with a situation-aware reward design, leading to a dynamic reward which allows the agent to learn situations which require controlled traffic rule exceptions. Our method is applicable to arbitrary RL models. We successfully demonstrate high completion rates of complex scenarios with recent model-based agents.
format Preprint
id arxiv_https___arxiv_org_abs_2402_04168
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Informed Reinforcement Learning for Situation-Aware Traffic Rule Exceptions
Bogdoll, Daniel
Qin, Jing
Nekolla, Moritz
Abouelazm, Ahmed
Joseph, Tim
Zöllner, J. Marius
Machine Learning
Computer Vision and Pattern Recognition
Robotics
Reinforcement Learning is a highly active research field with promising advancements. In the field of autonomous driving, however, often very simple scenarios are being examined. Common approaches use non-interpretable control commands as the action space and unstructured reward designs which lack structure. In this work, we introduce Informed Reinforcement Learning, where a structured rulebook is integrated as a knowledge source. We learn trajectories and asses them with a situation-aware reward design, leading to a dynamic reward which allows the agent to learn situations which require controlled traffic rule exceptions. Our method is applicable to arbitrary RL models. We successfully demonstrate high completion rates of complex scenarios with recent model-based agents.
title Informed Reinforcement Learning for Situation-Aware Traffic Rule Exceptions
topic Machine Learning
Computer Vision and Pattern Recognition
Robotics
url https://arxiv.org/abs/2402.04168