RigPI: Dynamic Parameter Identification of Rigid Body via VLM-Seeded Differentiable Simulation

2026-06-23

AI Summary

This work addresses the challenge of accurately identifying rigid-body inertial and friction parameters in real-world robotic manipulation, where sensor noise, modeling inaccuracies, and lack of prior knowledge hinder reliable estimation. To this end, the authors propose RigPI, a framework that integrates visual semantic priors, force/torque measurements, and motion observations to perform unconstrained dynamic parameter identification for multi-link rigid bodies within a differentiable simulation environment. A key innovation is the use of a vision-language model (VLM) to provide physically plausible initial parameter estimates and constrain the search space, complemented by a two-stage optimization strategy to enhance robustness. Experiments demonstrate that RigPI achieves accurate and stable parameter estimation on real objects with both revolute and prismatic joints, and successfully reproduces manipulation trajectories on a physical robot, validating the predictive fidelity of the identified dynamics.

Abstract

Accurate physical parameter identification of manipulated objects is fundamental to advanced robotic manipulation and the construction of faithful digital twins. However, acquiring physically consistent inertial and frictional properties from real-world interactions remains challenging due to sensing noise, modeling errors, and limited prior knowledge. This paper presents \textbf{RigPI}, a systematic framework for identifying dynamic parameters of both unconstrained rigid bodies and multi-link rigid bodies during robot-object interaction. RigPI integrates vision-based semantic priors, force-torque measurements, and motion observations within a differentiable simulation pipeline. A vision-language model (VLM) provides informed initialization and a constrained search space, while gradient information from a differentiable physics simulator enables efficient and stable parameter refinement. The proposed two-stage optimization strategy alleviates sensitivity to noise and avoids physically implausible solutions. Extensive real-world experiments on objects with revolute and prismatic joints demonstrate that RigPI achieves accurate and stable parameter estimates, and successfully reproduces manipulation trajectories on a real robot with parameter-aware predictive validity. These results highlight the effectiveness and robustness of RigPI for real-world robotic system identification tasks.

Problem

• dynamic parameter identification
• rigid body
• robotic manipulation
• digital twins
• system identification

Innovation

• differentiable simulation
• vision-language model
• dynamic parameter identification
• rigid body dynamics
• robotic manipulation