Generating optimal control simulations of musculoskeletal movement using OpenSim and MATLAB

Area of Science:

• Biomechanics
• Computational modeling
• Musculoskeletal dynamics

Background:

• Computer modeling, simulation, and optimization are increasingly vital in biomechanics research.
• Dynamic optimizations are categorized as data-tracking or predictive, with predictive approaches underutilized in clinical applications.
• Existing tools like OpenSim excel at tracking simulations but lack robust predictive capabilities.

Purpose of the Study:

• To develop a flexible framework for generating predictive musculoskeletal simulations.
• To integrate MATLAB's computational power with OpenSim's modeling capabilities.
• To facilitate the application of predictive simulations to clinically relevant human movement problems.

Main Methods:

• Developed a framework combining MATLAB and OpenSim for predictive musculoskeletal simulations.
• Utilized direct collocation optimal control techniques.
• Employed the open-source IPOPT solver and MATLAB's fmincon solver for problem-solving.

Main Results:

• Successfully generated predictive simulations for cyclical and discrete movements using varying musculoskeletal models.
• IPOPT solver demonstrated significant performance advantages over fmincon, especially for larger problems, by exploiting constraint Jacobian sparsity.
• Computation times were reasonable, ranging from seconds to a few hours.

Conclusions:

• The presented framework offers a powerful and adaptable method for creating optimal control simulations of musculoskeletal movement.
• This advancement is expected to increase the adoption of predictive simulation in clinical research for mobility-limiting conditions.
• Enhanced software tools are crucial for advancing predictive musculoskeletal simulations in clinical biomechanics.