Automated vehicles lack natural communication channels with other road users, making external Human-Machine Interfaces (eHMIs) essential for conveying intent and maintaining trust in shared environments. However, most eHMI studies rely on developer-crafted message-action pairs, which are difficult to adapt to diverse and dynamic traffic contexts. A promising alternative is to use Large Language Models (LLMs) as action designers that generate context-conditioned eHMI actions, yet such designers lack perceptual verification and typically depend on fixed prompts or costly human-annotated feedback for improvement. We present See2Refine, a human-free, closed-loop framework that uses vision-language model (VLM) perceptual evaluation as automated visual feedback to improve an LLM-based eHMI action designer. Given a driving context and a candidate eHMI action, the VLM evaluates the perceived appropriateness of the action, and this feedback is used to iteratively revise the designer's outputs, enabling systematic refinement without human supervision. We evaluate our framework across three eHMI modalities (lightbar, eyes, and arm) and multiple LLM model sizes. Across settings, our framework consistently outperforms prompt-only LLM designers and manually specified baselines in both VLM-based metrics and human-subject evaluations. Results further indicate that the improvements generalize across modalities and that VLM evaluations are well aligned with human preferences, supporting the robustness and effectiveness of See2Refine for scalable action design.

翻译：自动驾驶车辆缺乏与其他道路使用者的自然通信渠道，使得外部人机界面（eHMI）成为在共享环境中传达意图和维持信任的关键。然而，大多数eHMI研究依赖于开发者预先设计的消息-动作对，这难以适应多样且动态的交通场景。一种有前景的替代方案是使用大型语言模型（LLM）作为动作设计器，生成上下文相关的eHMI动作，但此类设计器缺乏感知验证，通常依赖固定提示或昂贵的人工标注反馈进行改进。我们提出See2Refine，一种无需人工干预的闭环框架，利用视觉语言模型（VLM）感知评估作为自动化视觉反馈，改进基于LLM的eHMI动作设计器。给定驾驶上下文和候选eHMI动作，VLM评估该动作的感知适当性，并将此反馈用于迭代修正设计器的输出，从而在无需人工监督的情况下实现系统性优化。我们在三种eHMI模态（灯条、眼睛和手臂）及多种LLM模型规模下评估了该框架。在所有设置中，我们的框架在基于VLM的指标和人类受试者评估中均一致优于仅依赖提示的LLM设计器及手工指定的基线。结果进一步表明，改进效果可跨模态泛化，且VLM评估与人类偏好高度一致，验证了See2Refine在可扩展动作设计中的鲁棒性和有效性。