Learning from Planned Data to Improve Robotic Pick-And-Place Planning Efficiency
Liang Qin, Weiwei Wan, Jun Takahashi, Ryo Negishi, Masaki Matsushita, Kensuke Harada
AI summary
Problem
Traditional grasp selection methods evaluate pick and place feasibility separately, causing rapid computational overhead as candidate sets grow. Directly learning shared grasps in high-dimensional spaces also demands excessive training data.
Approach
The method decomposes shared grasp prediction into two independent per-pose feasibility evaluations using an Energy-Based Model, then combines their energy scores to identify valid grasps efficiently.
Key results
- Significantly reduces planning time compared to analytical baselines
- Maintains high success rates across varying candidate set sizes
- Generalizes effectively to unseen grasp poses and table heights
- Compositional formulation improves data efficiency over direct prediction methods
Why it matters
Enables scalable, real-time grasp selection for industrial robotic workflows without sacrificing reliability or requiring massive datasets.
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. The 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 varying grasps and table heights, given that variations fall within the learned distributions.