数字式声呐中的升采样率处理

The rising sample rate processing in digital sonar

升采样率处理是某些数字式声呐的必不可少的预处理.本文给出整倍数升采样率运算的基本原理和理论结果。指出,升采样率处理的好坏不仅取决于窗函数的长度,也还和输入信号的功率谱密度有关.为实现升采样率处理,可以采用几种不同方法,文中指出用输入样本补零和窗函数卷积的方法是有效的.在数字计算机上的模拟取得了良好的结果.这种方法便于用专用数字信号处理芯片实现,并且对输入、输出采样率的选取上有很宽容的余地.

Rising sample rate processing (RSRP) is a necessary pre-processing in some cases for digital sonar. The principle of RSRP, in which the sample ratio of output over input is an integer, is given and the theoretical result is described. It is shown that the performance of RSRP not only depends on the length of the window function but also on the spectrum of input signal. Several algorithms can be chosen to implement the RSRP. It is shown that the method of convolution . by using the zero-paded input signal and the window function is effective. The system simulation have shown a good agreement of theoretical analysis. The algorithm described in this paper is easy to implement by digital signal processing chips. It is also affordable in choosing the input output sampling rate.