Plug-and-Play Label Map Diffusion for Universal Goal-Oriented Navigation

2026-05-07

AI Summary

This work addresses the challenge of simultaneously constructing semantic maps and localizing unobserved targets in unknown environments for goal-oriented navigation. To this end, the authors propose a plug-and-play Label Map Diffusion model (PLMD), built upon denoising diffusion probabilistic models (DDPMs). PLMD efficiently completes bird’s-eye-view semantic maps and accurately localizes targets under partial observability by integrating obstacle priors and structural consistency constraints. The approach mitigates semantic inconsistency issues prevalent in existing methods that rely on fully observed maps, enabling coherent semantic extension into unexplored regions. Experimental results demonstrate that PLMD achieves state-of-the-art performance across three navigation benchmarks and can be seamlessly integrated into existing navigation systems that depend on semantic maps.

Abstract

In embodied vision, Goal-Oriented Navigation (GON) requires robots to locate a specific goal within an unexplored environment. The primary challenge of GON arises from the need to construct a Bird's-Eye-View (BEV) map to understand the environment while simultaneously localizing an unobserved goal. Existing map-based methods typically employ self-centered semantic maps, often facing challenges such as reliance on complete maps or inconsistent semantic association. To this end, we propose Plug-and-Play Label Map Diffusion (PLMD), which defines a novel map completion diffusion model based on Denoising Diffusion Probabilistic Models (DDPM). PLMD generates obstacle and semantic labels for unobserved regions through a diffusion-based completion process, thereby enabling goal localization even in partially observed environments. Moreover, it mitigates inconsistent semantic association by leveraging structural consistency between known and unknown obstacle layouts and integrating obstacle priors into the semantic denoising process. By substituting predicted labels for unobserved regions, robots can accurately localize the specified objects. Extensive experiments demonstrate that PLMD \textbf{(I)} effectively expands the region of unknown maps, \textbf{(II)} integrates seamlessly into existing navigation strategies that rely on semantic maps, \textbf{(III)} achieves state-of-the-art performance on three GON tasks.

Problem

• Goal-Oriented Navigation
• Bird's-Eye-View map
• semantic map
• map completion
• embodied vision

Innovation

• Label Map Diffusion
• Goal-Oriented Navigation
• Denoising Diffusion Probabilistic Models
• Semantic Map Completion
• Bird's-Eye-View Mapping