Probabilistic approach improves robot placements for better object viewing
FlipToSee: A Probabilistic Stable Placement Prior for Active Visual Exploration via Regrasping
Robotics
Summary
Robots often need to turn objects placed on tables to see hidden parts. The authors developed FlipToSee, a method that predicts how to place an object stably so the robot can view more surfaces without trying every option. Their approach models possible placements smartly and ranks the best ones for safety and visibility. Tests in simulation and on real robots showed high success rates, even for unfamiliar objects.
What this means in practice
- •For robotics engineers: Integrate stable placement predictions with grasp and motion planning to enhance robotic exploration tasks involving unknown objects.
- •For automated warehouse operators: Improve robot handling by efficiently identifying safe object reorientations to expose hidden surfaces during packaging or sorting.
Authors
Chang Shu, Sushil Samuel Dinesh, Shinkyu Park
Abstract
Active visual exploration of tabletop objects often requires reorienting an unknown resting object onto a different stable support face to expose occluded surfaces. To identify such placements without exhaustive physical search, we learn a probabilistic placement prior from a single-view point cloud. Stable placement prediction is inherently multimodal, and conventional 6-DoF regression introduces further ambiguity by modeling translation and in-plane yaw. We therefore propose FlipToSee, a probabilistic framework that removes this representational ambiguity by parameterizing placements as unit support normals on $S^2$ while modeling their multimodal conditional distribution via a von Mises--Fisher mixture density network. To decouple mode diversity from physical robustness, FlipToSee deterministically extracts a compact candidate set from the mixture components and applies robustness-aware reranking using an auxiliary head trained with candidate-aligned supervision. In simulation, FlipToSee achieves $98.4\%$ first-proposal success on in-distribution objects, $95.3\%$ on out-of-distribution shapes, and $90.0\%$ under zero-shot transfer to household YCB objects. We further demonstrate the learned placement prior on a physical robot by integrating it with grasp and motion planning for exploratory regrasping.