Skip to end of metadata
Go to start of metadata

You are viewing an old version of this page. View the current version.

Compare with Current View Page History

Version 1 Current »

PhD student of Physical Therapy from Kyoto University, Japan

Workshop Goals

Error rendering macro 'viewpdf' : Failed to find attachment with Name OkitaYusuke.pdf

 

What to analyze:
Gait of patients after thigh muscle resection for soft tissue sarcoma

Expecting impact:
The knowledge which would be got from this study enables effective training depending on which muscles were resected.
(We can also apply the result to other patient population with muscle dysfunction!)

The changing goal:

At the time of application: Making a good MATLAB script to compare the experimental data with the simulated data
At the beginning of the workshop: To succeed in estimating muscle forces by CMC, learning the tips of tuning parameters (the former second goal)
Day 1: To successfully convert the model to gait2392 model and estimate the muscle forces using Static Optimization (Back to model preapration and Scaling!)

Working records:

Before workshop

Prepare the MATLAB scripts
 - the one which enables us to easily compare the experimental data (kinematics and kinetics calculated by Vicon, and EMG) with IK, ID, residuals, muscle forces, and activations
 - the one which can adjust the mass of each segment according to the recommendation after RRA
 - the one which delete the muscles you want using a simple GUI

Analysis using the model and experimental data
- I could run from Scaling-IK-SO-IAA without considering the accuracy of the model, verifying the result of ID was consistent with the kinetic data calculated by VICON

Day 1

Model simplification from Arnold-Hamner model to gait2392 model
- It costs a lot of time to use the model with wrappings compared with the one without them
- Walking task doesn't need complicated wrappings since the range of motion during movent is relatively small.
- We can cite the paper (Anderson and Pandy, 2001) when we choose not CMC but Static Optimization for muscle force estimation
- We can use more time for improving the results by avoiding time-consuming analysis such as CMC with the wrapping models

 lesson: careful consideration in deciding the model and methodology for estimating muscle forces considering the goal and the task of the study

Unsuccessful Scaling: the knee flexed too much in the static trial.

 - There were some differences in marker positions especially in ankle markers between the two models.
    -> improved by repositioning the marker

Inacurate IK results: Excessive hip external rotation during the swing phase

Possible causes and solutions

  1. the leg might be too long
  back to Scaling -> changing the original positions of markers - ASIS and Knee -> better tracking of the Knee marker and reduced the rotation

  2. poor tracking due to the small marker weight on Knee
    increase the marker weight on Knee marker -> got better result with the limb-length change

  3. skin movement ?

  lesson: Don't be relieved even if you checked the sagittal and frontal joint movement; horizontal movement (hip rotation) is remained.

Bugs that I experienced:

Before workshop

 - unable to use the model some of whose muscles are 'disabled' for CMC or Static Optimization. It stops before optimization process
   -> solved by deleting muscles from the .osim file.

 

Day 1

Scaling

- unable to use the setting file for the previous version
    -> solved by making a setting file from the beginning

 - unable to finish the process when repeating it using the same model
   -> unsolved

 

 

  • No labels