PRESTO: Fast motion planning using diffusion models based on key-configuration environment representation

2024-09-24

arXiv.org

AI Summary

To address the challenges of frequent collisions and low efficiency in robot motion planning within narrow passages, this paper proposes a diffusion model framework grounded in critical-configuration-based environmental representation. Methodologically, we introduce a novel C-space-aware sparse critical-configuration representation, explicitly embedding smoothness and collision-free constraints into the diffusion model’s training objective. The model learns high-quality trajectory priors and further refines collision-prone segments via gradient-based regularization. Our key contribution is the first deep integration of diffusion models with geometric constraints in configuration space, enabling effective synergy between learned guidance and optimization-based correction. Experiments in narrow-corridor scenarios demonstrate that our approach significantly improves planning success rate and computational efficiency over state-of-the-art learning-based and classical planners, while generating trajectories that are highly smooth, strictly collision-free, and real-time feasible.

Technology Category

• Intelligent Robots: Motion and Path Planning
• Computer Vision: Diffusion Models for Vision
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems

Abstract

We introduce a learning-guided motion planning framework that generates seed trajectories using a diffusion model for trajectory optimization. Given a workspace, our method approximates the configuration space (C-space) obstacles through an environment representation consisting of a sparse set of task-related key configurations, which is then used as a conditioning input to the diffusion model. The diffusion model integrates regularization terms that encourage smooth, collision-free trajectories during training, and trajectory optimization refines the generated seed trajectories to correct any colliding segments. Our experimental results demonstrate that high-quality trajectory priors, learned through our C-space-grounded diffusion model, enable the efficient generation of collision-free trajectories in narrow-passage environments, outperforming previous learning- and planning-based baselines. Videos and additional materials can be found on the project page: https://kiwi-sherbet.github.io/PRESTO.

Problem

• Generates collision-free trajectories using diffusion models.
• Approximates configuration space obstacles with key configurations.
• Improves motion planning efficiency in narrow-passage environments.

Innovation

• Diffusion model for trajectory optimization
• Sparse key-configuration environment representation
• Regularization for smooth collision-free trajectories