Drive As You Like: Multi-Head Diffusion with Reinforcement Learning for Personalized Driving
Year: 2025
Type: preprint
Abstract: Personalized autonomous driving requires understanding the driving needs implicit in user requests. This is challenging when driving demonstrations lack style annotations and behavioral adaptation risks degrading planning quality. We propose M-Diffusion Planner with a dual-decoupling design: separating intent understanding from trajectory planning, and general planning capability learning from strategy alignment. A large language model infers driving needs and selects a strategy, which is executed by the corresponding head of a shared diffusion planner. Following imitation pretraining, strategy-specific rewards specialize individual heads without style annotations. We innovatively adapt Group Relative Policy Optimization using surrogate policy-ratio estimation and constrained head-specific updates to mitigate degradation of pretrained planning capabilities. Open-loop and closed-loop evaluations on nuPlan demonstrate distinct strategy-aligned behaviors and competitive planning performance. Further experiments assess intent-to-strategy mapping and collision risk, and analyze the effects of reward design and optimization on strategy alignment and planning quality.
Source: arXiv (Cornell University)
Institutions: Monash University Malaysia, University of Utah, Monash University
Language: English
Subfield: Automotive Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7077907799
Cites:
Cited by:
Related to: