Performance Analysis of a Mass-Spring-Damper Deformable Linear Object Model in Robotic Simulation Frameworks

2025-04-18

AI Summary

Addressing the challenges of high-fidelity dynamic modeling of deformable linear objects (DLOs)—such as cables and wires—in robotic simulation, and the poor sim-to-real transfer of manipulation policies, this work proposes a physics-informed approach based on the mass-spring-damper (MSD) model. We systematically quantify, for the first time, the impact of key parameter perturbations on DLO dynamical responses and introduce a force-tactile data synthesis method integrated with domain randomization (DR). Experiments conducted in both Isaac Sim and Gazebo demonstrate that DR significantly enhances model robustness against uncertainties in material and geometric parameters, while improving cross-simulator generalization. This study empirically validates DR’s efficacy for DLO simulation and establishes a reproducible, scalable paradigm for synthetic force-tactile data generation—providing a reliable foundation for training simulation-based cable manipulation policies.

Technology Category

• Intelligent Robots: Manipulation
• Search and Optimization: Sampling/Simulation-based Search
• Planning, Routing, and Scheduling: Replanning and Plan Repair

Application Category

• Web Mining and Content Analysis: Web data generation and simulation
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

The modelling of Deformable Linear Objects (DLOs) such as cables, wires, and strings presents significant challenges due to their flexible and deformable nature. In robotics, accurately simulating the dynamic behavior of DLOs is essential for automating tasks like wire handling and assembly. The presented study is a preliminary analysis aimed at force data collection through domain randomization (DR) for training a robot in simulation, using a Mass-Spring-Damper (MSD) system as the reference model. The study aims to assess the impact of model parameter variations on DLO dynamics, using Isaac Sim and Gazebo to validate the applicability of DR technique in these scenarios.

Problem

• Modeling flexible DLOs like cables for robotic tasks
• Assessing parameter variations on DLO dynamics
• Validating domain randomization in simulation frameworks

Innovation

• Mass-Spring-Damper model for DLO simulation
• Domain randomization for force data collection
• Validation using Isaac Sim and Gazebo