Align2Act: Instruction-Tuned Models for Human-Aligned Autonomous Driving
This addresses the problem of ensuring autonomous vehicles behave in human-like and safe ways, representing a domain-specific incremental improvement over prior LLM-based methods.
The paper tackles the challenge of aligning motion planning in autonomous driving with human behavior by proposing Align2Act, a framework that uses instruction-tuned LLMs with structured reasoning, achieving an open-loop score of 85.17 and closed-loop scores up to 70.31 on the nuPlan benchmark.
Motion planning in complex scenarios is a core challenge in autonomous driving. Conventional methods apply predefined rules or learn from driving data to generate trajectories, while recent approaches leverage large language models (LLMs) for decision-making. However, it remains unclear whether LLMs truly capture human driving logic. We propose Align2Act, a motion planning framework that transforms instruction-tuned LLMs into interpretable planners aligned with human behavior. We derive structured driving instructions based on human reasoning patterns (e.g., anticipate hazards, yield at intersections) and traffic rules (e.g., stop at red lights, maintain lane boundaries). Our Align2ActChain module guides step-by-step reasoning to produce both an interpretable rationale and a safe trajectory. By fine-tuning LLaMA-2-7B with LoRA on one million scenarios from the nuPlan dataset, our method achieves an open-loop score of 85.17 and closed-loop scores of 70.31 (non-reactive) and 66.96 (reactive) on Test14-random. Unlike prior work focused on synthetic or open-loop settings, we demonstrate improved planning quality and human-likeness on the real-world nuPlan closed-loop benchmark. Ablation studies confirm that structured reasoning significantly improves performance over baseline LLM planners.