A Generalized Placeability Metric for Model-Free Unified Pick-and-Place Reasoning

2025-10-16

AI Summary

Real-world robotic manipulation faces challenges including absence of object priors (e.g., CAD models), high perception noise, and non-planar supporting surfaces. Method: We propose a shape-agnostic, generalized placeability metric framework that unifies grasp and placement reasoning via physics-based inference. Our approach automatically extracts multi-directional, non-planar support surfaces from raw point clouds and jointly models stability, graspability, and spatial clearance to generate collision-free, equilibrium-satisfying candidate poses; grasp quality is then scored in a pose-conditioned manner. Contribution/Results: On unseen real-world objects and complex supports, our method achieves stability prediction accuracy comparable to CAD-based baselines (orientation error <2.3°) and generates placements with superior physical plausibility and generalization over data-driven alternatives.

Technology Category

• Intelligent Robots: Manipulation
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Planning, Routing, and Scheduling: Replanning and Plan Repair

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• User Modeling, Personalization and Recommendation: Metrics for user behavior and evaluating success

Abstract

To reliably pick and place unknown objects under real-world sensing noise remains a challenging task, as existing methods rely on strong object priors (e.g., CAD models), or planar-support assumptions, limiting generalization and unified reasoning between grasping and placing. In this work, we introduce a generalized placeability metric that evaluates placement poses directly from noisy point clouds, without any shape priors. The metric jointly scores stability, graspability, and clearance. From raw geometry, we extract the support surfaces of the object to generate diverse candidates for multi-orientation placement and sample contacts that satisfy collision and stability constraints. By conditioning grasp scores on each candidate placement, our proposed method enables model-free unified pick-and-place reasoning and selects grasp-place pairs that lead to stable, collision-free placements. On unseen real objects and non-planar object supports, our metric delivers CAD-comparable accuracy in predicting stability loss and generally produces more physically plausible placements than learning-based predictors.

Problem

• Evaluates placement poses from noisy point clouds without shape priors
• Scores stability, graspability and clearance for unified pick-and-place reasoning
• Enables model-free manipulation on unseen objects and non-planar supports

Innovation

• Placeability metric evaluates poses from noisy point clouds
• Extracts support surfaces to generate multi-orientation placement candidates
• Conditions grasp scores on placements for unified pick-and-place reasoning