GUIDE: A Diffusion-Based Autonomous Robot Exploration Framework Using Global Graph Inference
Zijun Che, Yinghong Zhang, Shengyi Liang, Boyu Zhou, Jun Ma, Jinni ZHOU
AI summary
Problem
Existing exploration methods struggle to model unobserved spaces and plan globally efficient paths, often leading to redundant movements and suboptimal coverage due to myopic planning or high computational costs.
Approach
GUIDE constructs a unified global graph that merges observed data with predicted unexplored areas, filters predictions by regional reliability, and conditions a diffusion policy on this graph to generate stable, long-horizon exploration trajectories with fewer denoising steps.
Key results
- Up to 18.3% faster coverage completion
- 34.9% reduction in redundant movements
- Region-evaluated global graph filters unreliable spatial predictions
- Diffusion policy generates stable long-horizon trajectories with reduced computational overhead
Why it matters
Enables resource-constrained robots to achieve reliable, comprehensive coverage in complex indoor environments without heavy computational demands.
Abstract
Autonomous exploration in structured and com- plex indoor environments remains a challenging task, as existing methods often struggle to appropriately model unobserved space and plan globally efficient paths. To address these limita- tions, we propose GUIDE, a novel exploration framework that synergistically combines global graph inference with diffusion- based decision-making. We introduce a region-evaluation global graph representation that integrates both observed environ- mental data and predictions of unexplored areas, enhanced by a region-level evaluation mechanism to prioritize reliable structural inferences while discounting uncertain predictions. Building upon this enriched representation, a diffusion policy network generates stable, foresighted action sequences with significantly reduced denoising steps. Extensive simulations and real-world deployments demonstrate that GUIDE consistently outperforms state-of-the-art methods, achieving up to 18.3% faster coverage completion and a 34.9% reduction in redundant movements.