Informed Reinforcement Learning for Situation-Aware Traffic Rule Exceptions
arXiv:2402.04168 · doi:10.1109/ICRA57147.2024.10610842
Abstract
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.
Daniel Bogdoll and Jing Qin contributed equally. Accepted for publication at ICRA 2024
References in corpus (8)
- Automated Speed and Lane Change Decision Making using Deep Reinforcement Learning
- Mastering Diverse Domains through World Models
- Adaptive Behavior Generation for Autonomous Driving using Deep Reinforcement Learning with Compact Semantic States
- One Ontology to Rule Them All: Corner Case Scenarios for Autonomous Driving
- Trajectory Planning with Deep Reinforcement Learning in High-Level Action Spaces
- Knowledge Augmented Machine Learning with Applications in Autonomous Driving: A Survey
- A Fuzzy Approach to Qualification in Design Exploration for Autonomous Robots and Systems
- Safe Trajectory Planning Using Reinforcement Learning for Self Driving