Physics-Aware Object Position Learning

Updated 12 July 2026

Physics-aware object position learning integrates explicit physical constraints—such as non-penetration, gravity, and continuum dynamics—into pose estimation methods.
It spans various regimes, from local rigid 6D pose optimization using collision and gravity terms to global Lagrangian particle optimization for deformable objects.
Empirical results show that physics-aware refinements significantly improve 3D reconstruction accuracy and downstream manipulation performance in robotics and multi-view capture.

Markdown Report Issue Upgrade to Chat

References (8)

1.

PhysPose: Refining 6D Object Poses with Physical Constraints (2025)

2.

Improving Physics-Augmented Continuum Neural Radiance Field-Based Geometry-Agnostic System Identification with Lagrangian Particle Optimization (2024)

3.

High-Fidelity 4D Hand-Object Capture via Multi-View Spatiotemporal Tracking and Physics-Aware Gaussians (2026)

4.

Physics-Based Object 6D-Pose Estimation during Non-Prehensile Manipulation (2022)

5.

Learning Object Manipulation Skills from Video via Approximate Differentiable Physics (2022)

6.

Compact Multi-Object Placement Using Adjacency-Aware Reinforcement Learning (2024)

7.

Physics-informed Neural Time Fields for Prehensile Object Manipulation (2025)

8.

Learning Geometric Representations of Objects via Interaction (2023)

Topic to Video (Beta)

No one has generated a video about this topic yet.

Sign Up to Generate All Videos Subscribe on YouTube

Whiteboard

No one has generated a whiteboard explanation for this topic yet.

Sign Up to Generate

Follow Topic

Get notified by email when new papers are published related to Physics-Aware Object Position Learning.

Sign Up to Follow Topic by Email

Continue Learning