# Endowing Embodied Agents with Spatial Reasoning Capabilities for Vision-and-Language Navigation

Source: [Endowing Embodied Agents with Spatial Reasoning Capabilities for Vision-and-Language Navigation](<https://www.sota2.com/research/paper/68fa77eb1e209131cea76011>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-03-04T05:00:12.633Z

Authors: Qianqian Bai, Zhongpu Chen, Ling Luo, Huaming Du, Yuqian Lei, Ziyun Jiao

Published: 2025-04-09T02:19:22.000Z

## Abstract

Enhancing the spatial perception capabilities of mobile robots is crucial for achieving embodied Vision-and-Language Navigation (VLN). Although significant progress has been made in simulated environments, directly transferring these capabilities to real-world scenarios often results in severe hallucination phenomena, causing robots to lose effective spatial awareness. To address this issue, we propose BrainNav, a bio-inspired spatial cognitive navigation framework inspired by biological spatial cognition theories and cognitive map theory. BrainNav integrates dual-map (coordinate map and topological map) and dual-orientation (relative orientation and absolute orientation) strategies, enabling real-time navigation through dynamic scene capture and path planning. Its five core modules-Hippocampal Memory Hub, Visual Cortex Perception Engine, Parietal Spatial Constructor, Prefrontal Decision Center, and Cerebellar Motion Execution Unit-mimic biological cognitive functions to reduce spatial hallucinations and enhance adaptability. Validated in a zero-shot real-world lab environment using the Limo Pro robot, BrainNav, compatible with GPT-4, outperforms existing State-of-the-Art (SOTA) Vision-and-Language Navigation in Continuous Environments (VLN-CE) methods without fine-tuning.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Dynamic Obstacle Avoidance on Complex Instruction Tasks Avoidance](<https://www.sota2.com/research/sota/dynamic-obstacle-avoidance-on-complex-instruction-tasks-avoidance.md>): Dynamic Obstacle Avoidance; Complex Instruction Tasks Avoidance; Navigation Error (NE): 0.32.

- [Multi-step Navigation on Complex Instruction Tasks Multi-step](<https://www.sota2.com/research/sota/multi-step-navigation-on-complex-instruction-tasks-multi-step.md>): Multi-step Navigation; Complex Instruction Tasks Multi-step; Navigation Error (NE): 0.53.

- [Multi-target Search on Complex Instruction Tasks Multi-target](<https://www.sota2.com/research/sota/multi-target-search-on-complex-instruction-tasks-multi-target.md>): Multi-target Search; Complex Instruction Tasks Multi-target; Navigation Error (NE): 2.95.

- [Path Navigation on Single-command tasks environment](<https://www.sota2.com/research/sota/path-navigation-on-single-command-tasks-environment.md>): Path Navigation; Single-command tasks environment; Navigation Error (NE): 0.69.

- [Environment Interaction on Complex Instruction Tasks Interaction](<https://www.sota2.com/research/sota/environment-interaction-on-complex-instruction-tasks-interaction.md>): Environment Interaction; Complex Instruction Tasks Interaction; NE: 1.74.

- [Robot Navigation on Real-world Environments](<https://www.sota2.com/research/sota/robot-navigation-on-real-world-environments.md>): Robot navigation; Real-world Environments; BTR: 10.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2504.08806>)

[PDF](<https://arxiv.org/pdf/2504.08806.pdf>)
