Globally conservative high-order filters for large-eddy simulation and computational aero-acoustics

In computational aero-acoustics, large-eddy simulations (LES) or direct numerical simulations (DNS) are often employed for flow computations in the source region. As part of the numerical implementation or required modeling, explicit spatial filters are frequently employed. For instance, in LES spatial filters are employed in the formulation of various subgrid-scale (SGS) models such as the dynamic model or the variational multi-scale (VMS) Smagorinsky model; both in LES or DNS, spatial high-pass filters are often used to remove undesired grid-to-grid oscillations. Though these type of spatial filters adhere to local accuracy requirements, in practice, they often destroy global conservation properties in the presence of non-periodic boundaries conditions. This leads to the incorrect prediction of the flow properties near hard boundaries, such as walls. In the current work, we present globally conservative high-order accurate filters, which combine traditional filters at the internal points with one-sided conservative filters near the wall boundary. We test these filters to remove grid-to-grid oscillations both in a channel-flow case and in 2D cavity flow. We find that the use of a non-conservative filter leads to erroneous predictions of the skin friction in channel flows up to 30%. In the cavity-flow simulations, the use of non-conservative filters to remove grid-to-grid oscillations leads to important shifts in the Strouhal number of the dominant mode, and a change of the flow pattern inside the cavity. In all cases, the use of conservative high-order filter formulations to remove grid-to-grid oscillations lead to very satisfactory results. Finally, in our channel-flow test case, we also illustrate the importance of using conservative filters for the formulation of the VMS Smagorinsky model.     

A new neural network technique for the design of multilayered microwave shielded bandpass filters

In this work, we propose a novel technique based on neural networks, for the design of microwave filters in shielded printed technology. The technique uses radial basis function neural networks to represent the non linear relations between the quality factors and coupling coefficients, with the geometrical dimensions of the resonators. The radial basis function neural networks are employed for the first time in the design task of shielded printed filters, and permit a fast and precise operation with only a limited set of training data. Thanks to a new cascade configuration, a set of two neural networks provide the dimensions of the complete filter in a fast and accurate way. To improve the calculation of the geometrical dimensions, the neural networks can take as inputs both electrical parameters and physical dimensions computed by other neural networks. The neural network technique is combined with gradient based optimization methods to further improve the response of the filters. Results are presented to demonstrate the usefulness of the proposed technique for the design of practical microwave printed coupled line and hairpin filters. © 2008 Wiley Periodicals, Inc. Int J RF and Microwave CAE, 2009.     

Efficient temporal processing with biologically realistic dynamic synapses

Synapses play a central role in neural computation: the strengths of synaptic connections determine the function of a neural circuit. In conventional models of computation, synaptic strength is assumed to be a static quantity that changes only on the slow timescale of learning. In biological systems, however, synaptic strength undergoes dynamic modulation on rapid timescales through mechanisms such as short term facilitation and depression. Here we describe a general model of computation that exploits dynamic synapses, and use a backpropagation-like algorithm to adjust the synaptic parameters. We show that such gradient descent suffices to approximate a given quadratic filter by a rather small neural system with dynamic synapses. We also compare our network model to artificial neural networks designed for time series processing. Our numerical results are complemented by theoretical analyses which show that even with just a single hidden layer such networks can approximate a surprisingly large class of nonlinear filters: all filters that can be characterized by Volterra series. This result is robust with regard to various changes in the model for synaptic dynamics.     

Optimized orthonormal wavelet filters with improved frequency separation

This paper proposes an optimization-based method to design orthonormal wavelet filters with improved frequency separation. The proposed approach adopts a parameterization of orthogonal filter banks which ensures that the resulting wavelets have at least two vanishing moments. The filter parameters are adjusted by a numerical optimization algorithm in order to minimize a cost function associated to cut-off sharpness. In comparison with standard orthonormal filters, the proposed method is shown to provide better trade-off between frequency selectivity and time resolution. For illustration, the optimized filters are employed in an application example involving the use of a wavelet-packet system identification scheme. As a result, the identification errors are smaller than those obtained by using a non-optimized filter with the same length.     

Neural systems as nonlinear filters

Experimental data show that biological synapses behave quite differently from the symbolic synapses in all common artificial neural network models. Biological synapses are dynamic; their "weight" changes on a short timescale by several hundred percent in dependence of the past input to the synapse. In this article we address the question how this inherent synaptic dynamics (which should not be confused with long term learning) affects the computational power of a neural network. In particular, we analyze computations on temporal and spatiotemporal patterns, and we give a complete mathematical characterization of all filters that can be approximated by feedforward neural networks with dynamic synapses. It turns out that even with just a single hidden layer, such networks can approximate a very rich class of nonlinear filters: all filters that can be characterized by Volterra series. This result is robust with regard to various changes in the model for synaptic dynamics. Our characterization result provides for all nonlinear filters that are approximable by Volterra series a new complexity hierarchy related to the cost of implementing such filters in neural systems.     

用神经网络进行连续时间非线性系统建模的研究

在用神经网络进行系统建模时,建模误差的存在是难免的。为了减小这种误差,本文对连接时间非线性系统提出了一种新的神经网络辨识模型,它是由带有输入修正的神经网络和稳定滤波器组合而成。文中给出了权值的学习算法,即权值是根据辨识误差的投影算法来改变,证明了在一定条件下辨识误差的收敛性。     

Model updating of multistory shear buildings for simultaneous identification of mass,stiffness and damping matrices using two different soft-computing methods

In this research, two novel methods for simultaneous identification of mass–damping–stiffness of shear buildings are proposed. The first method presents a procedure to estimate the natural frequencies, modal damping ratios, and modal shapes of shear buildings from their forced vibration responses. To estimate the coefficient matrices of a state-space model, an auto-regressive exogenous excitation (ARX) model cooperating with a neural network concept is employed. The modal parameters of the structure are then evaluated from the eigenparameters of the coefficient matrix of the model. Finally, modal parameters are used to identify the physical/structural (i.e., mass, damping, and stiffness) matrices of the structure. In the second method, a direct strategy of physical/structural identification is developed from the dynamic responses of the structure without any eigenvalue analysis or optimization processes that are usually necessary in inverse problems. This method modifies the governing equations of motion based on relative responses of consecutive stories such that the new set of equations can be implemented in a cluster of artificial neural networks. The number of neural networks is equal to the number of degree-of-freedom of the structure. It is shown the noise effects may partially be eliminated by using high-order finite impulse response (FIR) filters in both methods. Finally, the feasibility and accuracy of the presented model updating methods are examined through numerical studies on multistory shear buildings using the simulated records with various noise levels. The excellent agreement of the obtained results with those of the finite element models shows the feasibility of the proposed methods.     

Online option price forecasting by using unscented Kalman filters and support vector machines

This study develops a hybrid model that combines unscented Kalman filters (UKFs) and support vector machines (SVMs) to implement an online option price predictor. In the hybrid model, the UKF is used to infer latent variables and make a prediction based on the Black–Scholes formula, while the SVM is employed to model the nonlinear residuals between the actual option prices and the UKF predictions. Taking option data traded in Taiwan Futures Exchange, this study examined the forecasting accuracy of the proposed model, and found that the new hybrid model is superior to pure SVM models or hybrid neural network models in terms of three types of options. This model can help investors for reducing their risk in online trading.     

复系数FIR数字滤波器的神经网络设计方法

神经网络是一种设计实系数FIR滤波器的有效算法,为了将该方法扩展到复数域,建立统一的基于神经网络的滤波器设计框架,文中提出了一种用多层神经网络设计任意幅频响应的复系数FIR数字滤波器的新算法,主要思想是将设计问题转化为实系数多层神经网络的训练问题,在实数域对幅频响应的平方误差函数的实部和虚部分别进行最小化,误差将收敛到全局最小点.实验结果表明,利用该算法设计的滤波器具有较小的幅频响应误差和群延迟误差.该算法能解决具有任意幅频响应和群延迟要求的问题,是一种有效的设计算法.     

Structure and properties of generalized adaptive neural filters forsignal enhancement

This article addresses the structure and properties of a new class of nonlinear adaptive filters called generalized adaptive neural filters (GANFs). Various properties, such as an upper bound of the mean absolute error of the filters, are analytically derived. Experimental results are presented to demonstrate the performance of the filters for signal and image enhancement. It is shown that GANFs not only extend the class of stack filters, but also have better performance in noise suppression.     

Shell segmentation requirements for numerical integration solutions

When numerical integration methods are used to solve shell equations, the exponential growth of solutions can lead to numerical difficulties in solving two-point boundary problems. This has led to the use of segmentation according to critical length criteria. The details of the calculations needed to solve a simple example have been examined to find how numerical difficulties arise. The effects of different precisions of calculation have been investigated. It is confirmed that for a single segment the critical length can be much greater than usually proposed. Reasons for this are suggested and it is shown that the same considerations can be applied to compound shells if certain suggested methods are employed.     

Neuro‐filters used to enhance the image quality of video displays

Abstract— Digital images can be affected by external factors. There are many types of noise which affect digital images. Image filtration is a basic method used to suppress such hindrances. The disadvantage of most filtration methods and hardware filters created on their behalf is their inability to react to changes in the input signal. The structure of the filters used for image processing is similar to the structure of a bi‐dimensional neural‐network matrix. Investigations have shown that a system with serial‐parallel filters of any degree of complexity can be created on the basis of the neural‐network matrices. Each neural‐network matrix layer acts as a separate neuro‐filter which can be trained and adapted to changes in the characteristics of the images. The neural‐network matrices allow for the creation of various types of linear and nonlinear filters, as well as combinations on the basis of a uniform structure. It allows for the design of a universal hardware neuro‐filter structure that can perform as different types of filters by means of loading the connectors weight. In our paper, we consider the realization of neuro‐filters based on a neural‐network matrix, which allows the processing of both static and moving images and increases the image sharpness, suppresses the noise, and detects movable objects in the processed image.     

未知输出反馈非线性时滞系统自适应神经网络跟踪控制

An adaptive output feedback neural network tracking controller is designed for a class of unknown output feedback nonlinear time-delay systems by using backstepping technique. Neural networks are used to approximate unknown time-delay functions. Delay-dependent filters are introduced for state estimation. The domination method is used to deal with the smooth time-delay basis functions. The adaptive bounding technique is employed to estimate the upper bound of the neural network reconstruction error. Based on Lyapunov-Krasoviskii functional, the semi-global uniform ultimate boundedness (SGUUB) of all the signals in the closed-loop system is proved. The arbitrary output tracking accuracy is achieved by tuning the design parameters and the neural node number. The feasibility is investigated by an illustrative simulation example.     

Adaptive Neural Tracking Control for Unknown Output Feedback Nonlinear Time-delay Systems

An adaptive output feedback neural network tracking controller is designed for a class of unknown output feedback nonlinear time-delay systems by using backstepping technique.Neural networks are used to approximate unknown time-delay functions.Delay-dependent filters are intro- duced for state estimation.The domination method is used to deal with the smooth time-delay basis functions.The adaptive bounding technique is employed to estimate the upper bound of the neural network reconstruction error.Based on Lyapunov-Krasoviskii functional,the semi-global uniform ultimate boundedness(SGUUB)of all the signals in the closed-loop system is proved.The arbitrary output tracking accuracy is achieved by tuning the design parameters and the neural node number. The feasibility is investigated by an illustrative simulation example.     

Morphological Connected Filtering on Viscous Lattices

This paper deals with the notion of connectivity in viscous lattices. In particular, a new family of morphological connected filters, called connected viscous filters is proposed. Connected viscous filters are completely determined by two criteria: size parameter and connectivity. The connection of these filters is defined on viscous lattices in such a way that they verify several properties of the traditionally known filters by reconstruction. Moreover, reconstruction algorithms used to implement filters by reconstruction can also be employed to implement these new filters. We also show that connected viscous filters have a behavior similar to filters with reconstruction criteria. The interest of these new connected filters is illustrated with different examples.     

Neural skyline filter for accelerating skyline search algorithms

The skyline search problem has been identified as one of the key problems in database research. None of the developed skyline search algorithms include the use of a filter to facilitate the search process. This paper proposes a novel modification involving the use of skyline filters to reduce the search space of a skyline problem by removing data points that cannot provide a viable skyline result. Three filters based on the concept of neural networks are proposed in this paper. The result is a reduction in execution time achieved through the reduction of the input tuples. The proposed filters may be used in conjunction with any existing skyline search algorithm. This is the first study to apply neural network technology to the skyline problem. Comprehensive simulation results demonstrate the effectiveness of the proposed skyline filtering system.     

Stabilization of switched neural networks with time‐varying delay via bumpless transfer control

This paper investigates the stabilization of switched neural networks with time‐varying delay. In order to overcome the drawback that the classical switching state feedback controller may generate the bumps at switching time, a new switching feedback controller which can smooth effectively the bumps is proposed. According to mode‐dependent average dwell time, new exponential stabilization results are deduced for switched neural networks under the proposed feedback controller. Based on a simple corollary, the procedures which are used to calculate the feedback control gain matrices are also obtained. Two simple numerical examples are employed to demonstrate the effectiveness of the proposed results.     

Adaptive output feedback control for nonlinear time-delay systems using neural network

This paper extends the adaptive neural network （NN） control approaches to a class of unknown output feedback nonlinear time-delay systems. An adaptive output feedback NN tracking controller is designed by backstepping technique. NNs are used to approximate unknown functions dependent on time delay, Delay-dependent filters are introduced for state estimation. The domination method is used to deal with the smooth time-delay basis functions. The adaptive bounding technique is employed to estimate the upper bound of the NN approximation errors. Based on Lyapunov- Krasovskii functional, the semi-global uniform ultimate boundedness of all the signals in the closed-loop system is proved, The feasibility is investigated by two illustrative simulation examples.