| Summary: | 碩士 === 國立交通大學 === 生醫工程研究所 === 104 === In recent years, biomedical sensing technology is developing rapidly. For the sake of “personal mobile healthcare”, now it is easier to achieve with the simple mobile devices, such as miniature bio-sensors or smart phones with bio-application. Although these non-linear and non-stationary bio-signals have been so easy to get, and how to more effectively analyze and applications are the focus of this thesis.
This these is based on Prof. John Kar-Kin, Zao and Chih-Kai, Lu’s patent "Feature Extraction from Human Gaiting Patterns using Principal Component Analysis and Multivariate Empirical Mode Decomposition" to analyze human motion triaxial acceleration signals, and then pick out the "Admissible Feature" to classification different Motions. We modified the Power Selecting Rule in this patent is used to select gait signals. We using three different ways to select gait signals: Mutual Information Selecting Rule, Instantaneous Frequency Selecting Rule and Power Spectral Density Selecting Rule, and to verify the accuracy of the selected rule by experiments. After the selecting the Gaiting signal and Impact signals, we can calculate the Plausible Feature, such as: gaiting signals' amplitude, frequency, and phase and peak points of impact signals, then we can select admissible feature from these plausible feature for classification different Motions.
|
|---|
