Stop Wandering: Efficient Vision-Language Navigation via Metacognitive Reasoning

2026-04-02

AI Summary

Existing training-free vision-language navigation methods often suffer from local oscillations and repetitive exploration due to a lack of metacognitive capabilities. This work proposes MetaNav, the first approach to integrate metacognitive reasoning into this task by constructing a persistent 3D semantic map for spatial memory, enabling history-aware path planning to avoid revisiting locations. When stagnation is detected, MetaNav leverages a large language model to generate corrective rules that dynamically adjust the navigation strategy. This framework achieves self-monitoring, diagnosis, and adaptive optimization of the navigation process, attaining state-of-the-art performance on GOAT-Bench, HM3D-OVON, and A-EQA while reducing vision-language model invocations by 20.7%.

Abstract

Training-free Vision-Language Navigation (VLN) agents powered by foundation models can follow instructions and explore 3D environments. However, existing approaches rely on greedy frontier selection and passive spatial memory, leading to inefficient behaviors such as local oscillation and redundant revisiting. We argue that this stems from a lack of metacognitive capabilities: the agent cannot monitor its exploration progress, diagnose strategy failures, or adapt accordingly. To address this, we propose MetaNav, a metacognitive navigation agent integrating spatial memory, history-aware planning, and reflective correction. Spatial memory builds a persistent 3D semantic map. History-aware planning penalizes revisiting to improve efficiency. Reflective correction detects stagnation and uses an LLM to generate corrective rules that guide future frontier selection. Experiments on GOAT-Bench, HM3D-OVON, and A-EQA show that MetaNav achieves state-of-the-art performance while reducing VLM queries by 20.7%, demonstrating that metacognitive reasoning significantly improves robustness and efficiency.

Problem

• Vision-Language Navigation
• Metacognitive Reasoning
• Spatial Memory
• Exploration Efficiency
• Foundation Models

Innovation

• metacognitive reasoning
• vision-language navigation
• spatial memory
• history-aware planning
• reflective correction