基于核主成分分析和支持向量机的飞机舱音信号识别  

Aircraft cockpit voice signal recognition based on kernel principal component analysis and SVM

作　　者：刘素京[1] 杨琳[2] 王从庆[1] 

机构地区：[1]南京航空航天大学自动化学院,南京210016 [2]中国民航总局航空安全技术中心,北京100028

出　　处：《东南大学学报（自然科学版）》2008年第A02期123-127,共5页Journal of Southeast University：Natural Science Edition

基　　金：国家自然科学基金资助项目(60776819)

摘　　要：为了提高飞机事故原因的调查准确性与实时性,提出了一种基于核主成分分析和支持向量机的舱音背景声识别方法.首先提取和分析了飞机驾驶舱话音记录器中所记录背景声信号的特征参数,然后分别以多项式核函数、sigmoid核函数和高斯核函数3种核函数作为内积,对3种核函数的降维特性进行了对比分析,最后将核方法与支持向量机结合,实现对舱音背景声的分类识别.实验结果表明:通过基于不同核函数的主成分分析方法与支持向量机的结合比较,确定以高斯核函数为内积的SVM分类方法具有较好的分类效果.In order to improve the accuracy and real-time performance of the aircraft accident research, a recognition method of background sound based on kernel principal component analysis and support vector machine （SVM） is proposed. First, feature coefficients of the background sound in the cockpit voice recorder are extracted and analyzed. Then by using polynomial kernel function, sigmoid kernel function and Gaussian kernel function as the inner product individually, the algorithm of the feature extraction based on kernel principal component analysis is researched, the reduced-order characteristics of three kinds of kernel function are compared and analyzed. Finally by combining kemel methods with SVM, the classification identification of the background sound is achieved. The simulation results show that, through the comparison of different kinds of kernel principal component analysis combined with support vector machine, the Gaussian kernel function method has preferable classification results.

关 键 词：特征提取 核主成分分析 支持向量机 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]