Graph-Guided Safe Diffuser: Topological Graph Guidance for Safe Diffusion Planning
2026-08-10 • Robotics
Robotics
AI summaryⓘ
The authors developed a new method, called Graph-Guided Safe Diffuser (G2SD), to help robots plan safe paths more reliably. Instead of changing paths on the fly, which can mess up the robot's movements, they first build a simple map (a graph) to plan high-level steps safely. Then, they use a diffusion model guided by this map to create smooth and feasible trajectories. Their tests showed that this approach greatly reduces collisions and improves reaching goals in navigation and movement tasks.
diffusion modelstrajectory planningtopological graphmanifoldkinematic feasibilitylatent graphrobot navigationsafety constraintsMaze2Dlocomotion
Authors
Nakgyu Yang, KwangBin Lee, SooJean Han
Abstract
Many diffusion-based planners enforce safety through inference-time guidance, but such interleaved trajectory deformations often degrade kinematic feasibility due to manifold rupture. We propose Graph-Guided Safe Diffuser (G2SD), a hierarchical framework that leverages a high-level topological graph planner to guide a low-level diffusion model. G2SD enforces safety at a structural level by abstracting the data manifold into a learned latent graph, on which high-level planning is performed. Continuous trajectories are generated by diffusion planners, which are conditioned on the graph node representations selected by the high-level planner. Theoretical analyses demonstrate conditions under which manifold rupture occurs in diffusion planners, and show that G2SD improves safety by reducing the constraint violation probability as the number of segments increases. Experiments demonstrate that G2SD substantially outperforms baselines, increasing goal-reaching rate without any collision from 40-50% to 98% in Maze2D navigation and also achieving superior task scores in locomotion.