Detecting and Tracking Human Movement Using Wearable Sensors

Wan, Yu

Detecting and Tracking Human Movement Using Wearable Sensors

May 7, 2015

New York Institute of Technology

Human locomotion
Medical informatics
Wearable computers

English

New York Institute of Technology 2015. All rights reserved.

PDF

Text

School of Engineering and Computing Sciences

Master of Science in Electrical and Computer Engineering

Ziqian (Cecilia) Dong

In this thesis, the problem of improving accuracy of detecting and tracking human motion using wearable sensors has been investigated. The study focuses on understanding movements disorders such as slowness of gait and freezing of gait for patients with Parkinson’s Disease. Adaptive fractal analysis is applied to interpret the time series data to distinguish healthy subjects and the subjects with PD and the movements from PD patients with or without freeze. This study used different analysis tools such as phase, hurst exponent and power spectral density to study the time series data to differentiate different activities. The features of the data sets is investigated by applying principal component analysis (PCA) and high frequency occurred features to reduce feature dimension. The estimation results were further evaluated with and without cross validation.