基于最小方差无偏估计的目标定位算法研究  

Research of Target Localization Algorithm Based on MVU

作　　者：陈泓宇[1] 孙超[1] 向龙凤[1] 

机构地区：[1]西北工业大学声学工程研究所,陕西西安710072

出　　处：《电声技术》2012年第1期39-42,共4页Audio Engineering

摘　　要：广义互相关时延估计方法(GCC)由于具有计算简单和易于实现的特点,因此在实际的被动声源定位系统中得到了广泛应用。随后出现了不同加权函数的GCC定位算法,但是在不同环境中各个加权函数的选择不同,时延估计误差不同。提出了基于最小方差无偏估计(MVU)的时延估计算法,对多种加权函数下的时延估计建立数学模型,然后用最小方差估计出最优的时延估计值。仿真实验表明,这种方法使估计精度明显提高。The GCC （Generalized Cross Correlation） method is widely used in a practical passive acoustic source localiza- tion systems because of its simple algorithm and easy realization. However, there are different weighting functions based on the GCC localization algorithm when applied to different environments, resulting different errors in time delay estimates. The time delay estimation algorithm based on MVU （Minimum Variance Unbiased）is put forword in this paper. The mathemati- cal model of the time delay estimation of various weighting functions is established, and the optimal value of time delay is es- timated by using the minimum variance. The results of simulation experiment clearly demonstrate that this method is superior to individual weighting of the GCC method.

关 键 词：时延估计 声源定位 广义互相关 最小方差无偏估计 

分 类 号：TN912.3[电子电信—通信与信息系统]