ANALYSIS AND MODELING OF LOCOMOTION - PROJECT 3 - README

Team members

• Albrecht Alice
• Juanico Alice
• Testa Laura

Data story

You can find the entire data story of this project on the complete report, called AML_Project3_datastory.pdf in the respository.

Structure of the repository

To allow the compiling of the given code it's necessary to have the same respository organization then indicate below.

The repository is composed of 4 folders:

• data: composed of 2 subfolders:
  • Healthy_data: which contains the raw data of the healthy patient for 3 conditions.
  • SCI Human: which contains the raw data of the spinal cord injured patient for 3 conditions.
• processed_data: where the process files of the 6 datasets will be stored.
• results: where the PCA results will be stored.
• scripts: which contains the following coding files:
  • gait_cycle_extraction.m: function that preprocess kinematics and extract the gait cycles.
  • gait_cycle_stickplots.m: function that allow a great visualization of one gait cycle per condition.
  • EMG_preprocessing.m: function that preprocess EMG data to obtain linear envelopes.
  • kinematic_features_extraction.m: function taht extract all the features from kinematics data preprocessed.
  • EMG_features_extraction.m: function that extract every all the features from EMG data preprocessed.
  • process_file.m: function that perform the complete processing of one file by calling all predefined functions.
  • main.m:

Setup

• To perform the entire analysis you just have to run the main.m file.
• We let you the possibility to show the processing plots by modfing in the main.m file at line 2: get_plots = 'False' into get_plots = 'True' The value is False by default to make the compiling of the code quicker.

Methods

To go through the complete analysis we performed multiple steps:

• Gait cycle extraction: Kinematics data are used to extract the gait cycles.
• EMG pre-processing: the EMG data is prepared for the feature extraction.
• EMG, kinematics features extraction: a list of specific features are extracted from the experimental data for each gait cycle.
• Principal Component Analysis (PCA): PCA is performed on the extracted features to compare injured and healthy locomotion.
• Visualization: results are visualized to better compare the locomotions.
• Statistical tests: statistical tests are performed to assess the critical parameters for a healthy locomotion