Loading...

Neuromuscular Modeling of Motor Control for Human Squat-to-Stand Motion Using Intelligent Algorithms

Mohammadi, Mahdi | 2024

0 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 57326 (08)
  4. University: Sharif University of Technology
  5. Department: Mechanical Engineering
  6. Advisor(s): Sayyaadi, Hassan
  7. Abstract:
  8. Motion strategies result from the management of priorities by the central nervous system in response to complex interactions with the peripheral nervous system, the musculoskeletal system, and the surrounding environment. The use of physics-based predictive neuro-mechanical simulations can play an important role in discovering and studying these complexities. With the development of detailed musculoskeletal models and neural models that include the control principles of the higher layers of the nervous system, it is possible to synthesize natural human movements, such as rising from a squat, without the need for experimental data and in novel conditions and environments. This can contribute to progress and reduce costs in the development of wearable technologies and rehabilitation devices. In the present research, the design of a neuro-musculoskeletal model for the predictive simulation of the squat-to-stand motion and the extraction of motion strategies related to it have been investigated. First, the requirements of a proper musculoskeletal model for the forward simulation of the squat-to-stand motion were examined, and based on these requirements, the model was prepared. Next, the problem of rising from a squat position was formulated using the musculoskeletal model as a direct collocation problem, and an appropriate cost function specific to the movement was designed for this trajectory optimization method. Subsequently, an algorithm was implemented to facilitate the tuning of the weight vector of the cost function and to obtain personalized cost functions with the help of experimental motion data. Finally, an experiment with different scenarios was designed and implemented to evaluate the performance of the designed framework. The results demonstrated high accuracy in tracking the kinematics, with an average RMSE of 0.075 radians for the joint angles across the four tested scenarios, compared to 0.203 radians in a similar study with two scenarios. Additionally, the model successfully covered a wide range of motion trajectories and succeed in extracting motion strategies
  9. Keywords:
  10. Musculoskeletal Modeling ; Squat Posture ; Optimal Control ; Inverse Optimal Control ; Squat-to-Stand Motion ; Neuromechanical Simulation

 Digital Object List

 Bookmark