Empirically-based Multibody Dynamics for Modeling the Human Body Musculoskeletal System

TL;DR

This paper presents a new empirically-based multibody dynamics framework for modeling the human musculoskeletal system, simplifying equations of motion and enabling realistic simulations of biological joints and muscle mechanics.

Contribution

It introduces a novel formalism using experimental data and matrix calculus to derive ordinary differential equations for musculoskeletal modeling, improving computational efficiency over DAE-based methods.

Findings

Successfully simulated human shoulder mechanism with empirically-derived constraints

Applied formalism to classical multibody systems like four-bar and five-bar linkages

Derived a closed-form formula for the musculotendon moment arm matrix

Abstract

This study introduces a novel approach for deriving the governing equations of the musculoskeletal system in the human body. The proposed formalism offers a framework to effectively incorporate the kinematic characteristics of biological joints and the complexities of kinematic chains into the differential equations of motion. This approach, known as "Empirically-based multibody dynamics," relies on experimental data pertaining to the skeletal system. To establish the formulations, a novel calculus of matrix-valued functions is employed. In contrast to alternative methods that utilize Differential-Algebraic Equations (DAEs), the current approach provides the governing equations of the musculoskeletal system in the form of ordinary differential equations, simplifying the computational process. This formalism shall be employed to simulate a generic constrained multibody system with holonomic constraints. The proposed formalism is applied to simulate various conventional multibody systems, including the four-bar linkage, five-bar linkage, and Andrew's squeezer mechanism. Additionally, the human shoulder mechanism is simulated, incorporating empirically-derived constraints to capture the shoulder rhythm. Furthermore, a novel method for deriving a closed-form formula for the moment arm matrix of a general musculotendon model is presented in this work.