SAIN: Structure-Aware Interactive Navigation with Active Dialogue Grounding for Mobile Robot

2026-08-10

AI Summary

This work addresses the challenge of mobile robots accurately localizing targets under ambiguous or incomplete natural language instructions. To this end, the authors propose the SAIN framework, which, for the first time, compiles human dialog feedback into a persistent structured navigation state—comprising target evidence, corridor memory, and candidate object labels—and jointly optimizes exploration and goal-approach behaviors through a unified policy network. Notably, SAIN requires no task-specific training and enables zero-shot long-horizon interactive navigation. Evaluated on the VL-LN IIGN benchmark, the method improves the success rate from 20.2% to 25.4% and increases the path-length-weighted success rate from 13.07 to 14.17, significantly outperforming existing dialog-based navigation baselines.

Abstract

Most existing vision-language navigation tasks assume that instructions are complete and unambiguous. However, real-world robots often encounter natural human instructions that are ambiguous, underspecified, or incomplete, requiring them to resolve such uncertainties through active questioning. Interactive Instance Goal Navigation (IIGN) requires an embodied agent to find the specific instance under an ambiguous category-level instruction through active dialogue. However, existing dialogue-enabled methods often consume oracle answers as transient textual context for immediate decisions, rather than persistent spatial or object-centric structured state. We present SAIN, a zero-shot framework that turns active dialogue into persistent navigation state. Instead of consuming oracle answers as one-step text hints, SAIN compiles them into target evidence, route-level corridor memory, and object-candidate labels. These states are stored in structured value, room, graph, and object memories, then consumed by a unified policy for frontier ranking and final target approach. On the VL-LN IIGN benchmark, SAIN improves SR from 20.2 to 25.4 and SPL from 13.07 to 14.17 over the strongest reported dialogue-enabled baseline, while requiring no task-specific policy training. The results support dialogue-to-state conversion as an effective zero-shot mechanism for long-horizon interactive instance navigation. Project website: https://zorattc.github.io/SAIN/

Problem

• vision-language navigation
• interactive navigation
• ambiguous instructions
• active dialogue
• instance goal navigation

Innovation

• interactive navigation
• active dialogue grounding
• structured memory
• zero-shot learning
• embodied AI