[Paper Review] A MATLAB Toolbox for Hybrid Rigid Soft Robots Based on the Geometric Variable Strain Approach

Anup Teejo Mathew, Ikhlas Ben Hmida+3 more|arXiv (Cornell University)|Jul 12, 2021

Soft Robotics and Applications4 citations

TL;DR

This paper presents SoRoSim, a MATLAB toolbox that uses the geometric variable-strain (GVS) approach based on Cosserat rod theory to enable accurate, fast, and computationally efficient static and dynamic modeling of soft, rigid, and hybrid robotic systems. The toolbox achieves real-time or faster-than-real-time simulation performance by minimizing degrees of freedom through geometric exactness, validated against literature and demonstrated across diverse robotic configurations including open-, branched-, and closed-chain systems with various actuators and loading conditions.

ABSTRACT

Soft robotics has been a trending topic within the robotics community for almost two decades. However, available tools for the modeling and analysis of soft robots are still limited. This paper introduces a user-friendly MATLAB toolbox, Soft Robot Simulator (SoRoSim), that integrates the Geometric Variable Strain (GVS) model of Cosserat rods to facilitate the static and dynamic analysis of soft, rigid, or hybrid robotic systems. We present a brief overview of the design and structure of the toolbox and validate it by comparing its results with those published in the literature. To highlight the toolbox's potential to efficiently model, simulate, optimize, and control various robotic systems, we demonstrate four sample applications. The demonstrated applications explore different actuator and external loading conditions of single-, branched-, open-, and closed-chain robotic systems. We think that the soft-robotics research community will significantly benefit from the SoRoSim toolbox for a wide variety of applications.

Motivation & Objective

• To address the lack of user-friendly, accurate, and computationally efficient tools for modeling and analyzing soft and hybrid robotic systems.
• To bridge the gap between traditional rigid robot modeling and soft robotics by extending geometrically exact methods to hybrid systems.
• To provide a unified, intuitive MATLAB-based framework for simulation, optimization, and control of soft robotic systems with minimal user-specific coding.
• To enable real-time or faster-than-real-time simulation for applications such as model-based control and design optimization.

Proposed method

• The toolbox implements the geometric variable-strain (GVS) formulation derived from Cosserat rod theory, enabling geometrically exact modeling of soft links as 1D continua.
• It uses a nested Gaussian quadrature scheme to numerically solve the GVS equations, improving computational efficiency and accuracy.
• The system is structured with user-friendly GUIs and MATLAB workspace integration for defining links, joints, degrees of freedom, constraints, and external/actuation forces.
• It supports customizable external and actuation force functions through editable MATLAB files, enabling integration with optimization and control toolboxes.
• The framework handles open-, branched-, and closed-chain robotic systems with both rigid and soft links, using minimal degrees of freedom through strain-based parametrization.
• It leverages MATLAB’s ecosystem, including the Optimization Toolbox and user-defined functions, for seamless extension to control and design tasks.

Experimental results

Research questions

• RQ1 Can a MATLAB-based toolbox using the GVS approach achieve high accuracy and computational efficiency in simulating soft and hybrid robotic systems?
• RQ2 How does the performance of the GVS-based approach compare to traditional FEM and lumped mass methods in terms of speed and degrees of freedom?
• RQ3 To what extent can the toolbox simulate diverse robotic configurations, including branched and closed-chain systems, under various actuation and loading conditions?
• RQ4 Can the toolbox support real-time or faster-than-real-time simulation for model-based control applications?
• RQ5 How well does the toolbox reproduce results from existing literature and numerical studies?

Key findings

• The SoRoSim toolbox achieves real-time or faster-than-real-time simulation performance across all tested examples, with simulation runtimes (ts) consistently lower than the real-time duration (tmax).
• The toolbox demonstrates high accuracy by successfully validating its results against published literature and numerical studies, confirming the reliability of the GVS formulation.
• The use of geometrically exact Cosserat rod modeling with variable strain reduces the number of degrees of freedom required, leading to faster and more efficient simulations compared to FEM and lumped mass approaches.
• The toolbox supports a wide range of robotic configurations, including single-, branched-, open-, and closed-chain systems, with flexible actuation and external loading setups.
• The integration with MATLAB toolboxes and user-defined functions enables seamless extension to optimization and control tasks, enhancing its utility for research and engineering applications.
• Performance benchmarks show that actuator strength computation is up to 2.5× faster than real time, enabling potential use in model-based inverse dynamic control.