权益分类	功能权益	普通用户	{{item.name}}会员
{{category.name}}	{{benefitItem.name}}

A Study on Auto-Regressive and Moving-Average Lattice Digital Filters

自回归和移动平均格子数字滤波器的研究

基本信息

批准号：
60550241
负责人：
NAGAI Nobuo
金额：
$ 1.6万
依托单位：
Research Institute of Applied Electricity, Hokkaido University
依托单位国家：
日本
项目类别：
Grant-in-Aid for General Scientific Research (C)
财政年份：
1985
资助国家：
日本
起止时间：
1985 至 1987
项目状态：
已结题

项目摘要

1. For the input and output signals of a stochastic process system, the transfer function is obtained by spectrum estimation or linear prediction. In this reseearch, we have developed Model Identification System (MIS) to estimate time varying parameters of ARMA model with an adaptive method. MIS is characterized by orthogonal relationships among prediction errors, new criterion for an approximation and a stability discrimination obtained by normalizing the prediction errors. By these properties, we can easily design the ARMA digital lattice filter with an arebitrary AR order and an arbitrary MA order. We also have proposed an algorithm possible of parallel processing for AR and ARMA modeling. By the algorithm, we can synthesize ARMA digital lattice filter faster.2. We apple the circuit theory for the synthesis of ARMA digital lattice filter which is suitable for adaptive digital signal processing. That is, we have shown that the distributed-constant circuit, which is an analog circuit, is equivalent to the Wave Digital Filter, and that they are equivalent to ARMA digital lattice filter. By the equivalence, we have denoted the synthesis method of ARMA digital lattice filters which match to the given inpulse responses and autocorrelation coefficients, or which are approximated with the least square algorithm.3. We consider the properties of a complex coefficient circuit by using the equivalent relation between the analog circuit and the digital circuit. In synthesizing the analog and digital circuits with the scattering matrix, we use reflection coefficients. There are two kinds of reflection coefficients for the complex circuit, i.e. equal matching and conjugate matching ones. We have introduced a pseudo-transmission-line suitable for the conjugate matching, and we show that the passive complex transfer function can be synthesized with the Darlington synthesis by using the pseudo-transmission-lines.

1.对于随机过程系统的输入和输出信号，传递函数可以通过谱估计或线性预测得到。本文提出了一种自适应估计阿尔马模型时变参数的模型辨识系统。MIS的特点是预测误差之间的正交关系，一个新的准则的近似和稳定性的判断，通过归一化的预测误差。利用这些性质，我们可以方便地设计任意AR阶和任意MA阶的阿尔马数字格型滤波器。提出了AR和阿尔马建模并行处理的算法。通过该算法，可以更快地综合出阿尔马数字格型滤波器.应用电路理论综合了适用于自适应数字信号处理的阿尔马数字格型滤波器。也就是说，我们已经证明，分布常数电路，这是一个模拟电路，是等效的波数字滤波器，他们是等效的阿尔马数字格型滤波器。通过等价性，给出了与给定的脉冲响应和自相关系数相匹配的阿尔马数字格型滤波器的综合方法，以及用最小二乘法近似的数字格型滤波器的综合方法.利用模拟电路与数字电路的等效关系，研究复系数电路的性质。在用散射矩阵综合模拟和数字电路时，我们使用反射系数。复电路的反射系数有两种，即等匹配反射系数和共轭匹配反射系数。本文介绍了一种适用于共轭匹配的伪传输线，并证明了利用伪传输线可以用达林顿综合法综合无源复传递函数。