Robotic system ranks grasp options to improve placement success rates
Execution-Aware Pre-Execution Ranking for Grasp-Conditioned Robotic Placement
Robotics
Summary
Robots need to pick up objects and put them down without knocking things over or getting stuck. The authors found that even when a spot to place an object looks right, the way a robot grabs it can make the placement tricky to execute. They built a model that scores different grab-and-place choices before the robot tries them, helping it pick ones more likely to work. Testing showed this approach improves success rates and works quickly on a real robot without extra retraining.
What this means in practice
- •For industrial robot programmers: Use pre-execution ranking to select grasp and placement pairs that improve successful object placement in manufacturing lines.
- •For robotics integrators: Deploy grasp-conditioning scoring models without retraining on new robot arms to speed up setup and improve automation reliability.
- •For warehouse automation operators: Increase pick-and-place efficiency by filtering grasp-placement candidates ahead of planning, reducing failed motions and downtime.$Commercial implications: Enables more effective robotic item handling products by improving placement success, reducing errors and operational costs.
Authors
Tianyuan Liu, Rutherford Agbeshi Patamia, Benjamin Champion, Richard Dazeley, Akansel Cosgun
Abstract
A geometrically valid placement can still be difficult to execute because the selected grasp changes the required end-effector pose, collision geometry, and transport motion. Placement is formulated as a pre-execution ranking problem in which supplied grasp-placement candidates are scored before planning. The model combines a typed target-conditioned point cloud with three pose descriptors and hierarchical heads for planning success and execution success conditioned on planning. On a 30-object, 1,235-scene dataset with scene-group-held-out splits, three-seed top-1 success on covered test groups reaches 85.63 +/- 1.08% for joint selection and 79.84 +/- 0.16% for fixed-target ranking. For the designated frozen seed-42 checkpoint, top-1 success improves from 72.84% to 85.78% over full-pool cuMotion for joint ranking and from 59.65% to 79.67% for fixed-target ranking. Frozen transfer to xArm7/MoveIt requires no xArm-specific retraining. Across 27 locked cases, 13 complete end to end (48.15%). Of the 16 cases that pass Top-5 preflight and begin execution, 13 succeed (81.25%). Candidate-level deployment-feasibility prediction reaches 81.25% recall, 85.20% specificity, and 83.23% balanced accuracy.