Breakthrough progress in vision-based navigation through unknown environments has been achieved by using multimodal large language models (MLLMs). These models can plan a sequence of motions by evaluating the current view at each time step against the task and goal given to the agent. However, current zero-shot Vision-and-Language Navigation (VLN) agents powered by MLLMs still tend to drift off course, halt prematurely, and achieve low overall success rates. We propose Three-Step Nav to counteract these failures with a three-view protocol: First, "look forward" to extract global landmarks and sketch a coarse plan. Then, "look now" to align the current visual observation with the next sub-goal for fine-grained guidance. Finally, "look backward" audits the entire trajectory to correct accumulated drift before stopping. Requiring no gradient updates or task-specific fine-tuning, our planner drops into existing VLN pipelines with minimal overhead. Three-Step Nav achieves state-of-the-art zero-shot performance on the R2R-CE and RxR-CE dataset. Our code is available at https://github.com/ZoeyZheng0/3-step-Nav.

翻译：通过利用多模态大语言模型（MLLMs），在未知环境中的视觉导航已取得突破性进展。这类模型可基于当前视角，结合赋予智能体的任务与目标，规划出一系列运动序列。然而，当前基于MLLMs的零样本视觉-语言导航（VLN）智能体仍存在偏离路线、过早停止及整体成功率偏低等问题。我们提出三步导航（Three-Step Nav）方法，通过三视角协议应对这些缺陷：首先，"向前看"提取全局地标并绘制粗略规划；其次，"看当前"将实时视觉观测与下一子目标对齐，实现精细引导；最后，"向后看"审计完整轨迹，在停止前修正累积的漂移误差。本规划器无需梯度更新或任务特定微调，可无缝嵌入现有VLN流水线且计算开销极低。三步导航在R2R-CE与RxR-CE数据集上实现了零样本性能的突破性提升。我们的代码已开源至https://github.com/ZoeyZheng0/3-step-Nav。