一类动态非线性系统模型频率域参数估值—维纳及准维纳模型

本文给出估计非线性系统Wiener及Wiener-like模型参数的新方法。这种方法有以下几个方面:(1) 输入是多音频正弦信号;② 通过测量输出的稳态反应,得到系统的各阶Volterra核;(3) 根据频率域内各阶Volterra的相互关系可以分离出线性子系统的传递函数;④通过分离的线性传递函数,此类模型的参数估值可以简化为线性系统的参数估值问题。     

基于GFRF的弱非线性模拟电路特征提取

介绍一种基于GFRF分析的非线性模拟电路故障诊断方法,着重阐述这种故障诊断技术中电路特征的提取方案,重点解决了激励信号的设计,Volterra频域核测量等关键技术问题.利用Volterra频域核辩识非线性系统,确定电路的最高显著阶,设计合适的激励信号,用范德蒙特方法分离出电路的各阶响应,利用公式计算出Volterra频域核.     

非线性动态网络的离散算法及仿真

在非线性网络响应分析中,采用Volterra级数法可以导出与线性系统传递函数相似的非线性传递函数,能使非线性系统用线性化和系统化方法达到精确分析.文中给出了非线性网络响应的Volterra级数解的连续算式,为解决连续算式计算麻烦的问题,提出用方波脉冲技术处理用Volterra级数表示法描述的非线性网络响应与激励之间关系的一组广义卷积积分的迭加计算,从而得到非线性网络响应求解的Volterra级数解的离散算式.仿真表明该算法求出的非线性网络响应与真实模型曲线十分逼近,证明了它的有效性.     

非线性ROF系统的分析和建模

在分析ROF(radio over fiber)系统组成与传输非线性的基础上,提出了用Volterra泛函级数为非线性ROF系统建立模型.研究了非线性ROF系统的Volterra级数表述、最小均方和最小二乘两种确定Volterra级数核的自适应算法在非线性ROF系统建模中的应用.通过一个用Volterra泛函级数对非线性ROF系统建模的实例,比较了Volterra LMS和Volterra RLS算法在建模中的效果,证明了用Volterra级数对非线性ROF系统建模的有效性.     

非线性网络Volterra级数表示的代数求解算法

本文基于非线性网络Volterra级数表示的辅助代数方法,提出了非线性网络n阶Volterra核和n阶非线性传递函数的代数编列算法。求取了非线性网络的规范形、三角形、对称形n阶Volterra核及相应的n阶非线性传递函数,同时也降低了非线性网络Volterra级数表示编列算法的复杂性。     

基于反向传播算法神经网络的谐波源模型分析

采用反向传播算法神经网络建立稳态频域的谐波源模型。在模型中,各次谐波电流的幅值和相角与各次谐波电压的幅值和相角以及负荷特征参数的非线性映射关系通过一种新颖的反向传播算法网络进行建模。该网络的学习算法是并行的。算例计算表明,该模型具有训练时间少、精度高等优点,是谐波源建模的有效方法。     

基于神经网络的Volterra频域核辨识方法

针对目前Volterra频域核辨识方法复杂、精度不高等问题,提出一种基于神经网络的Volterra频域核辨识方法。首先选择多组频率基准确测量各阶Volterra频域核的幅值,利用BP神经网络可以任意逼近非线性函数的特点,针对不同阶Volterra频域核设计不同的神经网络模型,进行分阶辨识,最后通过一个非线性电路进行仿真验证。仿真结果表明,该方法可直接辨识频率范围内任意频率对应的Volterra频域核,过程简单、准确度高,易于工程实现。     

测量前三阶Volterra核的QMMP算法

本文讨论频率域测量非线性系统Volterra核的快速准确算法。应用Volterra泛函级数和离散傅里叶变换理论于非线性系统,本文给出非线性系统在离散频率域的模型。应用该模型并适当地选取输入正弦信号的频率,可以同时多点地测量非线性系统的前三阶核。     

偏置相关GaAs MESFET建模分析

通过对利用多组偏置条件下的Ｓ参数获得精确ＡａＡｓＭＥＳＰＥＴ器件非线性模型方法的讨论，提出了新的沟道电流和栅电容模型，并提取了ＤＣ和电容模型参数。实验结果表明该模型模拟值和测量值吻合很好。     

偏置相关GaAs MESFET建模分析

通过对利用多组偏置条件下的Ｓ参数获得精确ＧａＡｓ ＭＥＳＦＥＴ器件非线性模型方法的讨论，提出了新的沟道电流的栅电容模型，并提取了ＤＣ和电容模型参数。实验结果表明该模型模拟值和测量值吻合很好。     

Blind identification of quadratic nonlinear models using neuralnetworks with higher order cumulants

A novel approach to blindly estimate kernels of any discrete- and finite-extent quadratic models in higher order cumulants domain based on artificial neural networks is proposed in this paper. The input signal is assumed an unobservable independently identically, distributed random sequence which is viable for engineering practice. Because of the properties of the third-order cumulant functions, identifiability of the nonlinear model holds, even when the model output measurement is corrupted by a Gaussian random disturbance. The proposed approach enables a nonlinear relationship between model kernels and model output cumulants to be established by means of neural networks. The approximation ability of the neural network with the weights-decoupled extended Kalman filter training algorithm is then used to estimate the model parameters. Theoretical statements and simulation examples together with practical application to the train vibration signals modeling corroborate that the developed methodology is capable of providing a very promising way to identify truncated Volterra models blindly     

Adaptive propagation modelling using a hybrid neural networktechnique

A hybrid modelling approach for the prediction of terrestrial wave propagation is introduced. It is shown that the use of neural networks leads to highly adaptive models by the simultaneous representation of nonlinear dependencies of various environmental parameters. This flexible and computationally effective approach can be used for calibration and as an extension of conventional prediction models     

实现标量混沌信号同步控制的神经网络方法

本文给出了一种利用线性输出神经网络实现标量混沌信号同步控制的方法。该方法利用线性输出神经网络构造被控混沌系统的模型，并基于Lyapunov理论与非线性系统控制方法，设计出神经网络权值变化规律与非线性反馈控制器，使神经网络模型的标量输出能大范围同步于给定的标量混沌信号。理论分析与计算机模拟结果都证实了这种方法的有效性。     

非线性负载简化模型及其应用

为实现复杂网络电磁脉冲（electromagnetic pulse，EMP）耦合分析，针对非线性负载引起的分析效率和收敛性问题，提出了非线性负载精简建模方法，即将多项式表示的非线性关系用压控元件代替，指定器件动作时间表示开关响应时间，忽略温度等不必要参数，减少模型元器件个数等.采用该方法建立了气体放电管（gas discharge tube，GDT）和金属氧化物变阻器（metal oxide varistors，MOV）的精简SPICE（simulation program with integrated circuit emphasis）模型，并且用3种不同类型的脉冲激励端接非线性负载的传输线，进行了收敛性和仿真效率分析.结果表明，GDT和MOV的简化模型能够很好收敛，MOV的模型分析效率提高约30%，GDT的模型分析效率稍有劣化，但完全避免了理想开关元器件的使用，与实际器件的工作原理一致.这些建模方法具有较强的普适性，可以移植到其他非线性负载的建模上，例如与线缆耦合的精简计算模型相结合，从而提高超大系统的电磁兼容（electromagnetic compatibility，EMC）设计与评估效率.     

基于遗传退火算法的短波频率分配问题研究

针对短波chirp选频系统的通信网络频率分配问题,建立了合理的干扰模型。同时利用系统所提供的参数,设计了新的算法参数。仿真结果表明,该算法在干扰约束条件下,算法能使全网通信质量达到最优,并对参数选择操作进行比较,给出了就如何设置算法运行参数的合理建议。     

一种半导体激光自混合效应模型参数的测量方法

从半导体激光自混合干涉系统的一般模型公式出发,建立了一个非线性函数模型,利用非线性最小二乘法的曲线拟合,来估计半导体激光自混合效应的模型参数——线宽展宽因数和光反馈因数。经计算仿真分析表明,该算法有较好的收敛性,迭代次数少且参数估计精度高,是半导体激光自混合效应参数测量的一种有效方法。     

A Scattering Variable Approach to the Volterra Analysis of Nonlinear Systems

A new mathematical model is developed which extends Volterra series analysis of nonlinear systems with memory to high-frequency systems, including those containing linear distributed component devices. A generalized set of nonlinear scattering parameters is defined which can be used to describe power transfer and distortion in nonlinear multiports, and which reduce to the classical scattering parameters for linear networks. The methodology is based on Volterra functional series, and is most useful for the small-signal case where the response can be approximated by a finite number of terms of the series. Nonlinear scattering kernels, derived by extending the Volterra analysis, are simply related to previously developed nonlinear voltage and current Volterra kernels. For sinusoidal inputs nonlinear scattering parameters are defined which are shown to be particularly helpful when power relationships are studied. The principal applications are for microwave networks terminated in real-valued site reference impedances. To evaluate the average power dissipated in a load at some intermodulation frequency, the concept of nonlinear transducer gain is defined and shown to be proportional to the squared magnitude of a nonlinear scattering parameter. Examples are presented illustrating the analysis procedure for a tunnel diode reflection amplifier and for a linear lossless transmission line terminated by a non-linear network.     

Oil well diagnosis by sensing terminal characteristics of theinduction motor

Oil well diagnosis usually requires dedicated sensors placed on the surface and the bottom of the well. There is significant interest in identifying the characteristics of an oil well by using data from these sensors and neural networks for data processing. The purpose of this paper is to identify oil well parameters by measuring the terminal characteristics of the induction motor driving the pumpjack. Information about oil well properties is hidden in instantaneous power waveforms. The extraction of this information was done using neural networks. For the purpose of training neural networks, a complex model of the system, which included 25 differential equations, was developed. Successful application of neural networks was possible due to the proposed signal preprocessing which reduces thousands of measured data points into 20 scalar variables. The special input pattern transformation was used to enhance the power of the neural networks. Two training algorithms, originally developed by authors, were used in the learning process. The presented approach does not require special instrumentation and can be used on any oil well with a pump driven by an induction motor. The quality of the oil well could be monitored continuously and proper adjustments could be made. The approach may lead to significant savings in electrical energy, which is required to pump the oil     

一种基于UPF的公路网约束下的GMTI跟踪方法

文中提出了一种利用公路网先验知识提高GMTI跟踪效果的新算法。公路网中的目标跟踪问题实质上是一个向量空间受约束的非线性估计问题。利用公路网这个约束条件，建立了一种新的目标状态模型，并利用适合于非线性非高斯状态的UPF算法进行跟踪滤波。仿真实验表明该算法降低了粒子滤波的计算量，提高了目标跟踪的精度。     

New neural method for bias dependent noise modelling of microwave transistors

A new method for accurate determination of noise parameters of microwave transistors for various bias conditions is proposed in this paper. The proposed model consists of a transistor empirical noise model (modification of Pospieszalski’s noise model) and two artificial neural networks. With the aim to avoid extraction of the empirical model parameters for each bias point, an artificial neural network is used to introduce bias-dependence of the equivalent circuit parameters. Accuracy of such bias-dependent model is further improved by using an additional neural network aimed to correct the noise parameters’ values. The proposed modeling approach is exemplified by modelling of a MESFET device in packaged form. The noise parameters obtained by the simulation agree well with the measured data.