Computational approach for modeling body and head movements during locomotion.
Item
-
Title
-
Computational approach for modeling body and head movements during locomotion.
-
Identifier
-
AAI3159228
-
identifier
-
3159228
-
Creator
-
Kunin, Mikhail.
-
Contributor
-
Adviser: Theodore Raphan
-
Date
-
2005
-
Language
-
English
-
Publisher
-
City University of New York.
-
Subject
-
Computer Science
-
Abstract
-
In this thesis, the nature of the invariant points in space that govern head translation and rotation in three-dimensional space and the anatomical basis for the coordinate frames that represent the head rotation was investigated. An algorithm was developed for approximating the pivots and axes for roll, pitch and yaw rotations of the head based on 4 x 4 displacement matrices, which represent translation and rotation of a rigid body in three dimensions. Head configurations in space were sampled (Optotrak 3020, Northern Digital, Inc.) and converted to a sequence of displacement matrices, D0, D1, ... DN. From a window of matrices around the ith sample, Di-n, ... Di, ... D i+n, a filter was obtained, giving the axis-angle of the rotation. Incremental displacement matrices around the ith sample were used to obtain vectors in the direction of the rotation axis from the rotational part of the incremental displacement matrix. An average axis was obtained for the window of displacement matrices by varying k. Axes were computed as a function of time by incrementing i from n to N-n. Procedures were developed to accurately calibrate the OPTOTRAK data relative to landmarks associated with the body and head. Studies of voluntary head movement and those during locomotion showed that the derived estimator was stationary and the axes of rotation of the head formed an approximately orthogonal basis. Determination of the yaw rotation axis compared favorably with the orientation of dens of C2 as determined from MRI studies. The method was also applied to finding head fixation points (HFP) in three dimensions using a Karhunen-Loeve transformation. The eccentricity of the distribution increased with walking velocity and was confined principally to the yaw plane in space. When the head movement was projected into the sagittal plane of the head, gain and phase of head rotation relative to its translation was obtained and could be modeled as a second order system, reflecting the behavior of the vestibulo-collic reflex (VCR). This new approach to filtering of three-dimensional rotation and translation of rigid body data gave insight into the compensatory mechanisms that govern head rotation in three dimensions during straight walking.
-
Type
-
dissertation
-
Source
-
PQT Legacy CUNY.xlsx
-
degree
-
Ph.D.
