摘要
介绍了作者发明的一种新的广义相关算法．介绍了该算法引入的思路．分析了该算法优于传统相关算法的几个主要原因．该算法将相关分析和倒频谱技术有机地结合起来，融两种分析方法的优点于一身，使传统相关函数分析方法无法分辨的时延峰在该算法中也能清楚地呈现．实验和理论表明该算法有着广泛的应用前景．
In this thesis a new generalized correlation algorithm and its train of thought to be invented are described in detail and several main causes in comparison with traditional correlation analysis method in time delay estimation are analysed. The new algorithm merges the advantages of two techniques:Cepstrum and Correlation. It can detect the time delay peak that traditional method cann't distinguish from signals with strong background noise. It is shown from the experimental and theoretical results that the new method has wide application prospect.
出处
《同济大学学报（自然科学版）》
EI
CAS
CSCD
1996年第6期684-691,共8页
Journal of Tongji University:Natural Science
基金
上海市科委资助项目
中德合作项目
关键词
倒谱
广义相关
时延估计
声信号
Cepstrum: Generalized correlation
Time delay estimation
