摘要
针对多径环境下麦克风阵列时延估计问题,将归一化最小均方误差法(NLMS)与互关联法(CR)相结合,提出了NLMS-CR算法.对算法的结构及原理作了详细分析,并与传统的相位变换广义互相关法(GCCPHAT)、NLMS进行比较;同时,在不同的信噪比(SNR)与声源环境下验证算法的效能.仿真结果表明:在较严重的多径环境下,NLMS-CR的性能优于其它传统的算法;当信噪比较高且接收信号相关性较低时,NLMS-CR具有很高的时延估计准确率.
An algorithm named NLMS- CR,combining the normalized least mean square（ NLMS） with the cross relation（ CR）,is presented in this paper to cope with the problem of time delay estimation for microphone arrays in multipath environment. The algorithm is made by a detailed analysis about its structure and principle,then compared with the traditional phase transform generalized cross- correlation（ GCC- PHAT） and NLMS. Simultaneously,the performance of the algorithm is verified on different SNR conditions and in different acoustic source environment. Simulations have shown that NLMS- CR is superior to other conventional algorithms in severe multipath environment. When SNR is high and the correlation of the received signals is low,NLMS- CR has a very high accuracy rate of time delay estimation.
出处
《嘉应学院学报》
2016年第8期43-46,共4页
Journal of Jiaying University
基金
江西省科学技术创新项目(GJJ12255)
广东省普通高校特色创新项目(2014KTSCX212)
关键词
多径环境
麦克风阵列
时延估计
算法
通信性能
multipath environment
microphone arrays
time delay estimation
algorithm
communication performance
