基于模糊神经网络的微波器件建模

通过模糊数据曲线来确定模糊规则,从而只需要根据系统的输入输出数据来建立网络最佳结构,并对BP算法进行了改进以加快网络模型的训练速度.在此基础上,用建立的模糊神经网络对微波无源器件共面波导的电容参数来进行提取.实验证明,模糊数据曲线确定的网络结构是接近最佳的,在训练速度和精度方面,效果优于神经网络建模算法.     

一种新型模糊神经网络结构确定的研究

本文给出了生种改进的模糊神经网络模型。这种模糊神经网络结构的建立只需要根据系统的输入输出数据来得到，且含义非常明晰。通常网络神经元的个数据根据经验来推，而我们通过糊数据曲线来确定网络模型，具有更高的准确度和优化性。     

图像目标区域定位模糊法实现

提出一种将传统的阈值法和模糊规则结合的图像分割方法，阈值法简便、快速、易行，模糊规则可以处理那些模棱两可的不确定的数据，两者结合的图像分割效果较自适应阈值法好，模糊规则由对实便图像的学习获得。     

基于模糊—神经网络混合系统的图象分割方法

本文提出了一种基于模糊－神经网络混合系统的图象分割方法。它可以利用人的经验知识和神经人样本数据中学习知识的能力，得到性能良好的模糊规则，并且可以通过神经网络结构实现模糊推理。     

一种神经网络模糊控制器的合成

论述了一种神经网络模糊控制器的合成方法，提出了单元-单元映象方法及利用模糊神经网络确定模糊中心，识别模型规则，调节隶属函数参数。仿真试验表明该模糊控制器相比PID控制器具有较好的性能。     

区间估计的FWNN及其区间学习算法

本文提出了一种基于子波的模糊神经网络，用以处理数据的区间估计，其网络结构基于模糊化的子波神经网络，在学习算法中，我们引入了区间学习算法的概念，使训练后的网络的输出为一个区间，其具体实现采用ＬＰ法，它同时提供了确定子波系数的另一种方法。     

神经网络自动生成模糊系统

在模糊系统的生成过程中，最主要的任务是隶属函数和模糊规则的提取和调整，但用传统方法，其工作量往往随变量数的增长而爆炸性地增加。为了解决这一问题，本文提出了一种新颖的，利用神经网络来自动地撮模糊系统 隶属函数和规则。     

基于标准阈值α和适用性阈值β的模糊分类

提出使用标准阈值α和适用性阈值β（α∈[0,1]且β∈[0,1]），从模糊分类问题的模糊化数据中产生模糊规则。用这个方法处理“星期六早晨问题”，比现有方法，分类准确率更高，而产生模糊规则较少。     

基于新的AFCM算法的陀螺仪漂移预测

为克服模糊规则提取的盲目性和随机性,提出了一种基于新的自适应模糊C-均值聚类(AFCM)算法的T-S 模糊建模方法.首先利用减法聚类来确定聚类数目的上限和初始聚类中心,然后采用改进的模糊C-均值聚类(FCM).算法进一步优化聚类中心,最后通过聚类有效性评判方法自适应地确定规则数及聚类中心,同时改进的FCM算法也克服了野...     

基于聚类模糊神经网络的非线性电路故障诊断

提出了一种基于聚类算法和模糊神经网络的非线性模拟电路故障诊断方法。通过一个无监督的聚类算法自组织地确定模糊规则的数目并生成一个初始的故障诊断模糊规则库，构造了一类模糊神经网络，通过训练调整网络权值，使故障诊断模糊规则库的分类更加精确，并通过仿真实验验证了该方法的有效性。     

A three-component scattering model for polarimetric SAR data

An approach has been developed that involves the fit of a combination of three simple scattering mechanisms to polarimetric SAR observations. The mechanisms are canopy scatter from a cloud of randomly oriented dipoles, evenor double-bounce scatter from a pair of orthogonal surfaces with different dielectric constants and Bragg scatter from a moderately rough surface. This composite scattering model is used to describe the polarimetric backscatter from naturally occurring scatterers. The model is shown to describe the behavior of polarimetric backscatter from tropical rain forests quite well by applying it to data from NASA/Jet Propulsion Laboratory's (JPLs) airborne polarimetric synthetic aperture radar (AIRSAR) system. The model fit allows clear discrimination between flooded and nonflooded forest and between forested and deforested areas, for example. The model is also shown to be usable as a predictive tool to estimate the effects of forest inundation and disturbance on the fully polarimetric radar signature. An advantage of this model fit approach is that the scattering contributions from the three basic scattering mechanisms can be estimated for clusters of pixels in polarimetric SAR images. Furthermore, it is shown that the contributions of the three scattering mechanisms to the HH, HV, and VV backscatter can be calculated from the model fit. Finally, this model fit approach is justified as a simplification of more complicated scattering models, which require many inputs to solve the forward scattering problem     

一种新的多尺度模型GLSA及其人脸识别的应用

考虑图像的近似信息和细节信息,提出了推广的线性尺度自回归(GLSA)多尺度模型。首先利用该模型建立不同尺度图像间的映射关系;其次使用原始图像及其小波分解结果得到模型中的参数以确定这种映射关系;最后根据该映射关系由低分辨率图像估计高分辨率图像。将该模型用于人脸识别处理,通过比较被测试图像的模型参数和训练集图像的模型参数确定被测试图像的类别。实验结果表明,使用GLSA模型估计得到的图像更加接近目标图像,以该模型为基础的人脸识别系统对光照的鲁棒性较强。     

一种基于广义KL距离和几何曲率的模型选择准则

模型选择的目标就是识别产生给定数据的模型.通常模型的好坏由模型的泛化能力来度量,而泛化能力包含模型对给定数据的拟合度和模型复杂度两个方面.本文从信息几何的观点使用定义在流形上的广义KL距离来度量模型的拟合度;另一方面从微分几何的观点用曲率的概念来度量模型的内在复杂度;因此,拟合度和复杂度的表示都具有在参数变换下保持不变的特点.通过理论分析,我们证明了用于表示模型预测能力的未来残差与模型固有曲率的关系.由此提出了一种新的基于广义KL距离和曲率的模型选择准则KLCIC.该准则不仅考虑了样本大小、参数个数和函数形式等影响复杂度的因素,而且具有非常清晰直观的几何意义.实验结果表明该方法的有效性.     

Automatic generation of accurate low-order models for magnetic devices

A method to automatically create low order state-space models for magnetic devices is set forth. The method begins with geometrical and materials data from which the finite element analysis (FEA) is used to build a high order state-space model. It is shown that this model contains far more state variables than are necessary to achieve the nearly the same results in the frequency band of interest. It is shown that the full-order model can be reduced methodically without additional physical assumptions. The reduction method is applicable directly to magnetically linear systems, but applies to a wide range of power electronics applications and can be a basis for future nonlinear device modeling. The method reduced an example inductor model from 882 state variables to two. To confirm correct model derivation, the full-order model is checked against and agrees with experimental data to the extent that the FEA and materials data are accurate. The reduced-order model agrees very well with the full-order model, particularly in computing impedance.     

Use of non-Rayleigh statistics for the identification of tumors in ultrasonic B-scans of the breast

A model for the scattering of ultrasound from breast tissue is proposed. The model is based on the use of non-Rayleigh statistics, specifically the K distribution to describe the backscattered echo from the tissue. A multiparameter test based on this model has been designed to characterize the tissue. The data from the B-scan images of the breasts of 6 different patients were analyzed using this model. The results indicate that the non-Rayleigh statistics seem to be useful in characterizing and identifying malignant, benign, and normal tissue regions     

An analytical two-dimensional model for silicon MESFETs

A model that predicts small-geometry effects in Si MESFETs has been developed. It is based on a two-dimensional (2-D) analytical solution of Poisson's equation in the subthreshold regime that applies to the junction-isolated structure typical of silicon devices. The model is in excellent agreement with numerical simulations from the PISCES 2-D device analysis program. The analytical model provides the physical basis for a subthreshold current model for small-geometry MESFETs. A scaling scheme for MESFETs, derived from the analytical model, that predicts a minimum-acceptable gate length of 0.15 μm for these devices is proposed     

数字IC的射频电磁发射模型建模方法

针对传统方法无法直接得到芯片内核到外部引脚传输路径特性的问题,提出了一种简单有效的数字IC的电磁发射建模方法。该模型由内部活动模型和传输路径模型组成。利用嵌入式环形振荡器的直流响应和外源射频干扰效应,直接测量出芯片内核到外部引脚路径的电压传输率,有效抽取出传输模型结构。利用晶体管级Hspice仿真得到内部活动模型,获得了完整的电磁发射模型。在SMIC 130 nm 工艺的测试芯片上应用该模型,对比模型仿真结果与实物测量结果,验证了该模型建模方法的正确性。该方法可用于系统设计阶段,对数字IC的电磁发射进行预测,使系统能满足电磁发射的要求。     

Port Based Actor Model with Kahn Process Network Model and Decidable Dataflow Model

This paper proposes a transitional actor model from legacy code to decidable dataflow. In addition, the proposed actor model provides dynamic behavior and top down design with static analysis such as deadlock detection and buffer memory size computation by combining Kahn process network (KPN) model and decidable dataflow (DCDF) model. In the proposed model, each port can have its own model of computation, which is different from the existing actor based models, so that it is called port based actor (PBA) model. Each port has either Kahn process network model or decidable dataflow model. A port group for KPN ports is introduced to specify KPN ports that are internally related. The proposed port based actor model is a generalized actor model extended from constant rate dataflow with intermediate ports (CRDF-IP) in which through the intermediate ports, an actor can consume and produce samples arbitrary times per execution. The decomposition of a PBA graph into DCDF graphs allows to apply static analysis, scheduling, and code generation methods developed for DCDF model. This paper explains formal definitions and static analysis for PBA model. Moreover, scheduling and efficient code generation methods are also explained. To validate that the proposed model works, PBA model has been implemented and an H.263 video encoder algorithm is specified and synthesized in PBA model.     

A model for residual life prediction based on Brownian motion with an adaptive drift

A degradation model is presented in this paper for the prediction of the residual life using an adapted Brownian motion-based approach with a drifting parameter. This model differs from other Brownian motion-based approaches in that the drifting parameter of the degradation process is adapted to the history of monitored information. This adaptation is performed by Kalman filtering. We also use a threshold distribution instead of the usual single threshold line which is sometime difficult to obtain in practice. We demonstrate the model using some examples and show that the model performs reasonably well and has a better prediction ability than the standard Brownian motion-based model. The model is then fitted to the data generated from a simulator using the expectation-maximization algorithm. We also fit a standard Brownian motion-based model to the same data to compare the difference and performance. The result shows that the adapted model performs better in terms of certain test statistics and the total mean square errors.     

汉语连续语音识别中上下文相关的识别单元(三音子)的研究

本文详细研究了汉语语音识别中如何有效地建立上下文相关的识别单元,以解决连续语音之间的协同发音问题。