發(fā)布時間：2018-05-05 21:44

  本文選題：遞歸圖 + 遞歸定量分析　； 參考：《西安交通大學學報》2017年10期

【摘要】：為了準確分辨并識別不同人體的步態(tài)特征,提出采用遞歸圖和遞歸定量分析的方法,計算人體步態(tài)非線性時間序列的復雜度。首先利用互信息和偽鄰近法分別計算得到合適的延時時間和嵌入維數(shù),根據(jù)相空間重構(gòu)的原理將一維時間序列擴展到高維相空間中,獲得時間序列在高維空間中鄰近點的分布規(guī)律和運動特點。構(gòu)建了患有帕金森疾病的老年人、健康老年人和健康年輕人的步態(tài)信號遞歸圖,可以直觀定性分析和評估這3組人群的步態(tài)信號的空間分布程度,其中健康人群最復雜。采用遞歸定量分析,量化了人體步態(tài)的復雜度,結(jié)果表明,患有帕金森疾病的人群的步態(tài)復雜度最小,而且獨立樣本t檢驗顯示了3組人群的復雜度具有顯著的差異性。該方法簡單可行,可以準確地對不同年齡和帕金森疾病的人群進行分類識別,有利于人體健康監(jiān)測和診斷研究。
[Abstract]:In order to distinguish and identify the gait features of different human bodies, the method of recursion and recursive quantitative analysis is proposed to calculate the complexity of the human gait nonlinear time series. First, the appropriate time delay time and the embedding dimension are calculated by mutual information and pseudo proximity, and the one dimensional time sequence is based on the principle of phase space reconstruction. The sequence is extended to the high dimensional space, and the distribution and motion characteristics of the adjacent points in the high dimensional space are obtained. The gait signal recursion of the elderly, healthy and healthy young people with Parkinson disease can be analyzed and evaluated directly to analyze and evaluate the spatial distribution of the gait signals of the 3 groups of people. The complex degree of human gait was quantified by recursive quantitative analysis. The results showed that the gait complexity of the people with Parkinson's disease was the least, and the independent sample t test showed that the complexity of the 3 groups was significantly different. This method was simple and feasible, and it could be accurate to different ages and Parkinson's disease. Identifying and classifying sick people is beneficial to the research of human health monitoring and diagnosis.

【作者單位】： 西安交通大學機械制造系統(tǒng)工程國家重點實驗室;香港中文大學機械與自動化工程學系;
【基金】：國家自然科學基金資助項目(51575426,51611530547) 中央高�；究蒲袠I(yè)務費綜合交叉項目(xjj2016002)
【分類號】：R318

【相似文獻】

相關(guān)期刊論文 前1條

1 張瑞紅,金德聞,張濟川,王人成,馬琳;不同路況下正常步態(tài)特征研究[J];清華大學學報(自然科學版);2000年08期

，

本文編號：1849419