Exploiting Prior Knowledge in Preferential Learning of Individualized Autonomous Vehicle Driving Styles

2025-03-19

AI Summary

This work addresses passenger-preference-driven generation of personalized autonomous driving styles, aiming to learn a trajectory planning cost function that jointly optimizes ride comfort and individual preferences while avoiding parameter configurations that induce passenger discomfort. Method: We propose a prior-augmented Bayesian optimization framework: (i) a virtual decision-maker—constructed from real-world human driving data—serves as a structured prior to guide safe and efficient parameter exploration; and (ii) closed-loop evaluation is enabled via integration of model predictive control with driver behavior modeling. Contribution/Results: Experiments demonstrate that our approach significantly accelerates convergence, reduces sampling of uncomfortable driving styles, and rapidly identifies the user’s optimal preferred style within a limited number of human interactions. It effectively resolves the longstanding trade-off between safety and efficiency in preference learning for autonomous vehicles.

Technology Category

Search and Optimization: Sampling/Simulation-based SearchHumans and AI: Learning Human Values and PreferencesPlanning, Routing, and Scheduling: Learning for Planning and Scheduling

Application Category

User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalizationResponsible Web: Machine-in-the-loop, human agency and autonomySearch and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Trajectory planning for automated vehicles commonly employs optimization over a moving horizon - Model Predictive Control - where the cost function critically influences the resulting driving style. However, finding a suitable cost function that results in a driving style preferred by passengers remains an ongoing challenge. We employ preferential Bayesian optimization to learn the cost function by iteratively querying a passenger's preference. Due to increasing dimensionality of the parameter space, preference learning approaches might struggle to find a suitable optimum with a limited number of experiments and expose the passenger to discomfort when exploring the parameter space. We address these challenges by incorporating prior knowledge into the preferential Bayesian optimization framework. Our method constructs a virtual decision maker from real-world human driving data to guide parameter sampling. In a simulation experiment, we achieve faster convergence of the prior-knowledge-informed learning procedure compared to existing preferential Bayesian optimization approaches and reduce the number of inadequate driving styles sampled.

Problem

• Optimizing cost function for preferred autonomous vehicle driving styles.
• Incorporating prior knowledge to improve preferential Bayesian optimization.
• Reducing passenger discomfort and speeding up learning convergence.

Innovation

• Uses preferential Bayesian optimization for learning
• Incorporates prior knowledge from human driving data
• Constructs virtual decision maker for parameter sampling