Learning from Planned Data to Improve Robotic Pick-and-Place Planning Efficiency

TL;DR

This paper introduces a learning-based approach using an Energy-Based Model to predict shared grasps, significantly accelerating robotic pick-and-place planning by reducing computational overhead and improving generalization.

Contribution

The work presents a novel EBM framework for shared grasp prediction, enhancing efficiency and generalization in robotic pick-and-place tasks.

Findings

Improves grasp selection performance

Offers higher data efficiency

Generalizes well to unseen objects

Abstract

This work proposes a learning method to accelerate robotic pick-and-place planning by predicting shared grasps. Shared grasps are defined as grasp poses feasible to both the initial and goal object configurations in a pick-and-place task. Traditional analytical methods for solving shared grasps evaluate grasp candidates separately, leading to substantial computational overhead as the candidate set grows. To overcome the limitation, we introduce an Energy-Based Model (EBM) that predicts shared grasps by combining the energies of feasible grasps at both object poses. This formulation enables early identification of promising candidates and significantly reduces the search space. Experiments show that our method improves grasp selection performance, offers higher data efficiency, and generalizes well to unseen grasps and similarly shaped objects.