BoT-Drive: Hierarchical Behavior and Trajectory Planning for Autonomous Driving using POMDPs

2024-09-27

arXiv.org

AI Summary

Uncertainty in human driving behavior within dynamic road environments severely compromises the safety and reliability of autonomous vehicle planning. To address this, we propose a hierarchical behavior-trajectory co-planning framework formulated within a Partially Observable Markov Decision Process (POMDP), unifying driver intent recognition, implicit driving style inference, and ego-vehicle decision-making. Our key contributions are twofold: (1) we pioneer the use of a parametric driver model as both a latent-state (intent/style) estimator and a generator of the ego-vehicle’s executable action space; and (2) we introduce a behavior-constrained importance sampling mechanism to enable fine-grained trajectory optimization. Evaluated on real-world urban driving scenarios, our method achieves significantly higher collision avoidance rates and improved trajectory stability. It consistently outperforms conventional rule-based and end-to-end learning planners across multiple safety and efficiency metrics.

Technology Category

• Humans and AI: Human-Aware Planning and Behavior Prediction
• Planning, Routing, and Scheduling: Planning with Markov Models (MDPs, POMDPs)
• Multiagent Systems: Multiagent Planning

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• User Modeling, Personalization and Recommendation: Metrics for user behavior and evaluating success

Abstract

Uncertainties in dynamic road environments pose significant challenges for behavior and trajectory planning in autonomous driving. This paper introduces BoT-Drive, a planning algorithm that addresses uncertainties at both behavior and trajectory levels within a Partially Observable Markov Decision Process (POMDP) framework. BoT-Drive employs driver models to characterize unknown behavioral intentions and utilizes their model parameters to infer hidden driving styles. By also treating driver models as decision-making actions for the autonomous vehicle, BoT-Drive effectively tackles the exponential complexity inherent in POMDPs. To enhance safety and robustness, the planner further applies importance sampling to refine the driving trajectory conditioned on the planned high-level behavior. Evaluation on real-world data shows that BoT-Drive consistently outperforms both existing planning methods and learning-based methods in regular and complex urban driving scenes, demonstrating significant improvements in driving safety and reliability.

Problem

• Addresses uncertainties in dynamic road environments for autonomous driving
• Manages complexity of behavioral intentions and hidden driving styles
• Enhances safety through hierarchical planning and trajectory optimization

Innovation

• Hierarchical POMDP planning for autonomous driving
• Driver models infer hidden driving styles
• Trajectory optimization using importance sampling