Loading...

Dynamic Modeling and Simulation of Knee Joint Using an Artificial Neural Network Model

Amouzandeh, Reza | 2018

586 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 51431 (08)
  4. University: Sharif University of Technology
  5. Department: Mechanical Engineering
  6. Advisor(s): Sayadi, Hassan
  7. Abstract:
  8. One of the important challenges in gait analysis is to know the consequences and extent of the impact of an injury or failure and the outcome of any intervention. The purpose of this study was to identify and categorize three gait patterns of a healthy person, an individual with ruptured anterior cruciate ligament, and an individual with surgically reconstructed anterior cruciate ligament using knee joint kinematic information in normal walking. To do this, data from the knee joint curve during the gait cycle in sagittal, frontal and transverse planes were used as a neural network input. Four neural networks were designed to detect and isolate the three gait patterns from each other. The structure of all these networks had one hidden layer with ten neurons. In the first three neural networks, only one of the flexion-extension, abduction-adduction or internal-external rotation angles was used for training and evaluation of gait patterns. The fourth neural network used a combination of the angles of all three anatomical planes for training and evaluation. Trained neural networks with flexion-extension and abduction-adduction angles did not have a good accuracy for separating the three gait patterns. The accuracy of the trained neural network with rotation angle data was 82.2% and the final neural network accuracy was 91.2%. This accuracy was more than 93% in determining the gait pattern of an anterior cruciate ligament rupture. With these results, it can be claimed that it is possible to categorize these three patterns of gait by using knee joint kinematic information and the proper structure of the neural network
  9. Keywords:
  10. Knee Joint ; Anterior Cruciate Ligament (ACL) ; Artificial Neural Network ; Knee Kinematics ; Damage Identification ; Dynamic Modeling

 Digital Object List

 Bookmark

No TOC