Musculoskeletal Simulation Based Optimization of Rehabilitation Program


Biologist who study animal locomotion have long speculated the control, structure and actuation has been optimized by evolution for locomotion. In the class of vertebrates, the focus of this paper, such locomotion is derived from cyclic/periodic motions of locomotory appendages (arms, legs, fins, wings), implemented in the form of integrated neuromusculoskeletal systems. The suitable coordination and neural control of the contraction of the muscles attached to an underlying skeletal structure allows for realization of desired periodic motion-trajectories (gaits). It is noteworthy that antagonistic co-contraction of the highly redundant musculature, has also been shown to provide significant robustness to environmental disturbances (modulated force interactions). However, our focus is on steady state gait where the consensus opinion is that they have been optimized to reduce the overall effort. Our long-term interests are in systematically studying natural and robotic systems, from the view point of: (i) obtaining insights that will allow for better design of the robotic systems; and (ii) using the robotic counterpart as a scientific tool for studying hypotheses regarding the natural system. In particular, one commonly-discussed hypothesis is that expended effort is minimal when the locomotory appendages/propulsive structures are moving at steady-state in periodic gaits at their resonant frequencies. Ahlborn et al. showed that humans, and by extension all animals, maintain resonance during walking and running [1, 2]. In this paper, we will focus on bipedal walking as a prototypical task for the following reasons: (i) It possesses considerable richness in terms of complexity of underlying neuromusculoskeletal structure and control; (ii) considerable past work (experimental and mathematical) is available to help crosscheck the results; and (iii) most importantly, there is considerable interest currently in realizing bipedal robotic systems, where efficiency is extremely critical.

Various modeling, analysis and experimental approaches have been employed in the past to study this very important problem (shown above). However, in this work, we will hypothesize that under “normal” steady-state gait conditions, the neuromusculoskeletal system has been optimized so as to minimize the expended effort when it reaches resonance. To validate this hypothesis, we propose the use of a virtual prototyping/ simulation-based-design methodology using contemporary computational tools (musculoskeletal analysis coupled with a parametric sweep/optimization tool) to facilitate the systematic study initially of simplified pendulum models and subsequently full-fledged bipedal walking, completely within a virtual environment.


 Students Involved:

- Leng-Feng Lee, Ph.D. Student.



 Movies :


1. Musculoskeletal Simulation based Study of Biped Locomotion

- A video explain the idea, analysis, and results of this work, created for BIOROB 2008.

- Larger Screen: view it on YouTube:










 Related Publications - Conference Proceedings:


Lee, L-F, and Krovi, V. “Musculoskeletal Simulation of Optimal Gait Frequency in Biped and Human Locomotion,” Proceeding of IEEE BIOROB 2008, Scottsdale, Arizona, October 19-22, 2008. (Best Poster Award)



Sponsor: This project was funded by Research Foundation of State University of New York, National Science Foundation CAREER Award (IIS-0347653) and CNS-0751132.

Last Updated: August 28, 2010