Three-Step Nav: A Hierarchical Global-Local Planner for Zero-Shot Vision-and-Language Navigation

TL;DR

Three-Step Nav introduces a hierarchical, three-view planning protocol for zero-shot vision-and-language navigation, significantly improving success rates without additional training.

Contribution

It proposes a novel three-view hierarchical planning method that enhances zero-shot VLN performance by global landmark extraction, fine-grained alignment, and trajectory correction.

Findings

Achieves state-of-the-art zero-shot performance on R2R-CE and RxR-CE datasets.

Does not require gradient updates or task-specific fine-tuning.

Integrates seamlessly into existing VLN pipelines with minimal overhead.

Abstract

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.