Loading...

Feet Motion Pattern Recognition Based on Data from Swiss Ranger Camera and other Dynamical Sensors with Offering an Intelligent Data Bank in Motion Correction and Rehabilitation

Sharifi Kolarijani, Arman | 2011

314 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 41888 (08)
  4. University: Sharif University of Technology
  5. Department: Mechanical Engineering
  6. Advisor(s): Sayyaadi, Hassan
  7. Abstract:
  8. In this project, we use an off-line motion capture technique for walking pattern recognition and reconstruction, and apply a sensory data acquisition system for extracting knee and ankle joints angular movements. Our approach is a marker-based object tracking technique using only one camera and four markers as data acquisition system, and a neural network method is developed for marker detection. Different walking condition considered including: straight position, inclined position with 〖+5〗^° slope, and 〖-10〗^° slope. Using pattern obtained from above process, the BVH format of the motion is built for 3D reconstruction of the performer’s motion. Since the obtain data describe the motion of one leg, an intelligent algorithm is introduced to generate the other leg data from the one we have for full reconstruction of both legs. In order to validate the simulated motion, a comparative analysis is performed between simulated and real motion.Also, kinematic analyses are performed in different conditions for all joints, and in order to validate the results, a comparative analysis is performed between our approach’s results and the gait that is used in other references. Then, we compared angular and transitional kinematic characteristics of different joints in the different conditions, and propose some linguistic analyses of joints of these joints, in sagittal plane. And, we compare the results of these kinematics analyses with sensory data acquisition system we propose, which is a digital encoder. These comparison shows that both of this approach has similar results although in sensory data acquisition system, because of some limitation, we can not record some token characteristics of leg’s joints. Furthermore, it has been seen that major differences between inclined slope gaits and straight gaits are during stance phases of gait cycles
  9. Keywords:
  10. Image Processing ; Neural Network ; Simulation ; Gait Analysis ; Encoder ; Resilient Back-Propagation (RBP)Learning Algorithm ; Biovision Hierarchy (BVH)Format

 Digital Object List

 Bookmark

No TOC