Study on a Wavelet-Extracted-Eigenvector Based Representation of Time Series on Recognition of Astronaut_s Respiratory Intensity

This paper presents a time series´ representation by eigenvector sequence based on wavelet transform. This representation is applicable for recognizing astronaut´s respiratory intensity. By wavelet analysis, the time series of respiratory intensity is separated into noise and de-noised curve. Noise filtered effectively, the de-noised curve is smooth enough for facile subsection. According to such representation, this paper implements the similarity comparison and recognizes the character of respiratory intensity of astronaut in China manned space flight experiment.