NDPP-Grasp: Non-Differentiable Physical Plausibility Constraint-Guided Task-Oriented Dexterous Grasp Generation
For robotic grasping, this work addresses the decoupling of task alignment and physical plausibility in diffusion-based grasp generation, offering a practical method to incorporate non-differentiable constraints during generation.
The paper proposes a framework that integrates non-differentiable physical plausibility constraints directly into the denoising process of a task-aligned grasp diffusion model, improving grasp quality without post-generation refinement. Experiments show significant gains in physical plausibility and task success over baselines.
Task-oriented dexterous grasp generation aims to produce dexterous grasp poses that are both physically plausible and functionally suitable for specified manipulation tasks. Existing diffusion-based methods often address these two requirements in a decoupled manner: they first train a grasp diffusion model for task alignment and then rely on post-generation refinement to improve physical plausibility. However, this after-the-fact correction strategy applies physical plausibility guidance only once the grasp has already been generated, leaving the generation trajectory itself unguided by physical constraints and potentially leading to suboptimal grasps. To address this problem, we propose a novel framework that directly injects physical plausibility guidance into the denoising process of a task-aligned grasp diffusion model in a practical and effective manner, even when physical plausibility constraints are non-differentiable. This allows physical plausibility to shape grasp generation throughout denoising while preserving task alignment. Extensive experiments demonstrate the efficacy of our framework.