Finite Impulse Response Neural Network Research Articles

Hardware Implementation of FIR Neural Network for Applications in Time Series Data Prediction

Time series data prediction is used in several applications in the area of science and engineering. Time series prediction models have been implemented using statistical approaches, but recently, neural networks are being applied for times series prediction due to their inherent properties and capabilities. A variation of a standard neural network called as finite impulse response (FIR) neural network has proven to be highly successful in achieving higher degree of prediction accuracy when used over various time series prediction applications. These applications are time critical and involve huge amounts of computation that are slower when run on a general purpose processor and hence, a dedicated hardware is required. In this paper, authors present hardware implementation of an FIR neural network for applications in times series data prediction. The implementation is divided into (i) off-board, where the training algorithm and neural network configuration is implemented in Matrix Laboratory (MATLAB) and simulated with various benchmark time series data set and (ii) on-board, where the entire system is modeled in a hardware description language (HDL). The simulation experiment, hardware building blocks, the implementation framework, and the hardware design flow are discussed in this paper. The hardware resource utilization and timing information are also reported in the paper. DOI: http://dx.doi.org/10.11591/telkomnika.v14i1.7272

Downhole Pressure Estimation Using Committee Machines and Neural Networks

In gas-lifted oil wells the monitoring of downhole pressure plays an important role. However, the permanent downhole gauge (PDG) sensor often fails. Because maintenance or replacement of PDGs is usually unfeasible, soft-sensors are promising alternatives to monitor the downhole pressure in the case of sensor failure. In this paper, a data-driven soft-sensor is implemented to estimate the downhole pressure using committee machines composed by finite impulse response (FIR) neural networks. Experimental results in three real datasets of the same oil well indicate that the identified soft-sensor is able to predict the downhole pressure with satisfactory accuracy. The model input variables were selected by statistical tests which increased insight concerning such variables. Committee machines outperformed single-model soft-sensors on experimental data.

Common virtual wind speed sensors for wind farm based on finite impulse response neural network

Wind speed sensors of wind turbines has high fault rate,those faults may lead wind turbines into safety risks and energy production loss.However,many current methods of improving reliability of wind speed information face the challenges of high cost or high error.A virtual wind speed sensor based on spatial correlation was presented in this paper.Its key character is generating a downwind turbine's logic wind speed based on a special upwind turbine's real wind speed.The calculation mode of logic wind speed just need the outputs of the existed wind speed and direction sensors.A FIR(Finite Impulse Response) neural network computing model was proposed for deal with the complexity of logic wind speed calculating.Moreover,key technologies were discussed for building the virtual wind speed sensor system.The logic wind speed generated by virtual sensor can be used for providing reliable wind speed information input to a turbine controller.The approach presented in this paper is applicable to wind turbines of any type in wind farms.

A novel multiple-channel active noise control approach with neural secondary-path model for interior acoustic noise attenuation of railway train systems

Interior noise cancellation for railway train systems is an important means of enhancing passenger comfort and quality of service. This study proposes a novel active noise control (ANC) approach that uses an finite impulse response (IIR) filter and neural network techniques to effectively reduce interior noise. The authors construct a multiple-channel IIR filter module that is a linearly augmented framework with a generic IIR model to generate a primary control signal. A three-layer perceptron neural network is employed for establishing a secondary-path model to represent air channels among noise fields. Since the IIR module and neural network are connected in series, the output of an IIR filter is transferred forward to the neural model to generate a final ANC signal. A gradient descent optimisation-based learning algorithm is analytically derived for the optimal selection of the ANC parameter vectors. Moreover, re-estimation of partial parameter vectors in the ANC system is proposed for online learning. Sufficient stability conditions are derived for the proposed ANC system. Lastly, the authors present the results of a numerical study to test their ANC methodology with realistic interior noise measurement obtained from Korean railway trains.

Modeling frequency modulated responses of midbrain auditory neurons based on trigger features and artificial neural networks

Frequency modulation (FM) is an important building block of communication signals for animals and human. Attempts to predict the response of central neurons to FM sounds have not been very successful, though achieving successful results could bring insights regarding the underlying neural mechanisms. Here we proposed a new method to predict responses of FM-sensitive neurons in the auditory midbrain. First we recorded single unit responses in anesthetized rats using a random FM tone to construct their spectro-temporal receptive fields (STRFs). Training of neurons in the artificial neural network to respond to a second random FM tone was based on the temporal information derived from the STRF. Specifically, the time window covered by the presumed trigger feature and its delay time to spike occurrence were used to train a finite impulse response neural network (FIRNN) to respond to this random FM. Finally we tested the model performance in predicting the response to another similar FM stimuli (third random FM tone). We found good performance in predicting the time of responses if not also the response magnitudes. Furthermore, the weighting function of the FIRNN showed temporal ‘bumps’ suggesting temporal integration of synaptic inputs from different frequency laminae.This article is part of a Special Issue entitled: Neural Coding.

Adaptively Combined FIR and Functional Link Artificial Neural Network Equalizer for Nonlinear Communication Channel

This paper proposes a novel computational efficient adaptive nonlinear equalizer based on combination of finite impulse response (FIR) filter and functional link artificial neural network (CFFLANN) to compensate linear and nonlinear distortions in nonlinear communication channel. This convex nonlinear combination results in improving the speed while retaining the lower steady-state error. In addition, since the CFFLANN needs not the hidden layers, which exist in conventional neural-network-based equalizers, it exhibits a simpler structure than the traditional neural networks (NNs) and can require less computational burden during the training mode. Moreover, appropriate adaptation algorithm for the proposed equalizer is derived by the modified least mean square (MLMS). Results obtained from the simulations clearly show that the proposed equalizer using the MLMS algorithm can availably eliminate various intensity linear and nonlinear distortions, and be provided with better anti-jamming performance. Furthermore, comparisons of the mean squared error (MSE), the bit error rate (BER), and the effect of eigenvalue ratio (EVR) of input correlation matrix are presented.

Linear combination rule in genetic algorithm for optimization of finite impulse response neural network to predict natural chaotic time series

A finite impulse response neural network, with tap delay lines after each neuron in hidden layer, is used. Genetic algorithm with arithmetic decimal crossover and Roulette selection with normal probability mutation method with linear combination rule is used for optimization of FIR neural network. The method is applied for prediction of several important and benchmarks chaotic time series such as: geomagnetic activity index natural time series and famous Mackey–Glass time series. The results of simulations shows that applying dynamic neural models for modeling of highly nonlinear chaotic systems is more satisfactory with respect to feed forward neural networks. Likewise, global optimization method such as genetic algorithm is more efficient in comparison of nonlinear gradient based optimization methods like momentum term, conjugate gradient.

Multiresolution FIR neural-network-based learning algorithm applied to network traffic prediction

In this paper, a multiresolution finite-impulse-response (FIR) neural-network-based learning algorithm using the maximal overlap discrete wavelet transform (MODWT) is proposed. The multiresolution learning algorithm employs the analysis framework of wavelet theory, which decomposes a signal into wavelet coefficients and scaling coefficients. The translation-invariant property of the MODWT allows alignment of events in a multiresolution analysis with respect to the original time series and, therefore, preserving the integrity of some transient events. A learning algorithm is also derived for adapting the gain of the activation functions at each level of resolution. The proposed multiresolution FIR neural-network-based learning algorithm is applied to network traffic prediction (real-world aggregate Ethernet traffic data) with comparable results. These results indicate that the generalization ability of the FIR neural network is improved by the proposed multiresolution learning algorithm.

Neural networks as effective techniques in clinical management of patients: some case studies

In this paper, we present four examples of effective implementation of neural systems in the daily clinical practice. There are two main goals in this work; the first one is to show that neural networks are especially well-suited tools for solving different kind of medical/pharmaceutical problems, given the complex input output relationships and the few a priori knowledge about data distribution and variable relations. The second goal is to develop specific software applications, which enclose complex mathematical models, to clinicians; thus, the use of such models as decision support systems is facilitated. Four important pharmaceutical problems are considered in this study: identification of patients with potential risk of postchemotherapy emesis, classification of patients depending on their risk of digoxin intoxication, prediction of cyclosporine A through concentration and prediction of erythropoietin blood concentrations. The Multilayer Perceptron in classification problems and dynamic neural networks, such as the Elman recurrent neural network and the Finite Impulse Response neural network in prediction problems, have been used. Moreover, network ensembles of different kind of networks have been taken into account. Results show that neural networks are suitable tools for medical classification and prediction tasks, outperforming the mostly used methods in these problems (logistic regression and multivariate analysis).

Foetal ECG recovery using dynamic neural networks

Non-invasive electrocardiography has proven to be a very interesting method for obtaining information about the foetus state and thus to assure its well-being during pregnancy. One of the main applications in this field is foetal electrocardiogram (ECG) recovery by means of automatic methods. Evident problems found in the literature are the limited number of available registers, the lack of performance indicators, and the limited use of non-linear adaptive methods. In order to circumvent these problems, we first introduce the generation of synthetic registers and discuss the influence of different kinds of noise to the modelling. Second, a method which is based on numerical (correlation coefficient) and statistical (analysis of variance, ANOVA) measures allows us to select the best recovery model. Finally, finite impulse response (FIR) and gamma neural networks are included in the adaptive noise cancellation (ANC) scheme in order to provide highly non-linear, dynamic capabilities to the recovery model. Neural networks are benchmarked with classical adaptive methods such as the least mean squares (LMS) and the normalized LMS (NLMS) algorithms in simulated and real registers and some conclusions are drawn. For synthetic registers, the most determinant factor in the identification of the models is the foetal–maternal signal-to-noise ratio (SNR). In addition, as the electromyogram contribution becomes more relevant, neural networks clearly outperform the LMS-based algorithm. From the ANOVA test, we found statistical differences between LMS-based models and neural models when complex situations (high foetal–maternal and foetal-noise SNRs) were present. These conclusions were confirmed after doing robustness tests on synthetic registers, visual inspection of the recovered signals and calculation of the recognition rates of foetal R-peaks for real situations. Finally, the best compromise between model complexity and outcomes was provided by the FIR neural network. Both the methodology for selecting a model and the introduction of advanced neural models are the main contributions of this paper.

Responses of central auditory neurons modeled with finite impulse response (FIR) neural networks

To simulate central auditory responses to complex sounds, a computational model was implemented. It consists of a multi-scale classification process, and an artificial neural network composed of two modules of finite impulse response (FIR) neural networks connected to a maximum network. Electrical activities of single auditory neurons were recorded at the rat midbrain in response to a repetitive pseudo-random frequency modulated (FM) sound. The multi-scale classification process divides the training dataset into either strong or weak response using a multiple-scale Gaussian filter that based on response probability. Two modules of FIR neural network are then independently trained to model the two types of responses. This caters for the possible differences in neuronal circuitry and transmission delay. Their outputs are connected to a maximum network to generate the final output. After training, we use a different set of FM responses collected from the same neuron to test the performance of the model. Two criteria are adopted for assessment. One measures the matching of the modeled output to the actual output on a point-to-point basis. Another measures the matching of bulk responses between the two. Results show that the proposed model predicts the responses of central auditory neurons satisfactorily.

Optical finite impulse response neural networks

The finite impulse response neural network is described in detail. Different algorithms capable of temporal back-propagation are considered, including a novel modification to the conventional algorithm, called the delayed-feedback back-propagation algorithm. We present and analyze different optoelectronic processors making use of adaptive volume holograms and three-dimensional optical processing. Two single-layer architectures are presented: the input delay plane architecture and the output delay plane architecture. By combining them it is possible to implement both forward and backward propagation in two multi-layer architectures: the first making use of the conventional temporal back-propagation and the second making use of delayed-feedback back-propagation.

Non-Linear Unsteady Aerodynamic Response Approximation Using Multi-Layer Functionals

Non-linear functional representation of the aerodynamic response provides a convenient mathematical model for motion-induced unsteady transonic aerodynamic loads response, that accounts for both complex non-linearities and time-history effects. A recent development, based on functional approximation theory, has established a novel functional form; namely, the multi-layer functional. For a large class of non-linear dynamic systems, such multi-layer functional representations can be realised via finite impulse response (FIR) neural networks. Identification of an appropriate FIR neural network model is facilitated by means of a supervised training process in which a limited sample of system input-output data sets is presented to the temporal neural network. The present work describes a procedure for the systematic identification of parameterised neural network models of motion-induced unsteady transonic aerodynamic loads response. The training process is based on a conventional genetic algorithm to optimise the network architecture, combined with a simplified random search algorithm to update weight and bias values. Application of the scheme to representative transonic aerodynamic loads response data for a bidimensional airfoil executing finite-amplitude motion in transonic flow is used to demonstrate the feasibility of the approach. The approach is shown to furnish a satisfactory generalisation property to different motion histories over a range of Mach numbers in the transonic regime.

Time series prediction using artificial neural networks

In this work, artificial neural network (ANN) algorithms were used to predict time series of the oceanographic variables Southern Oscilation Index (SOI) and sea surface temperature anomaly (SSTA). The finite impulse response neural network (FIRNN) was applied to data obtained from the NOAA. In order to determine the most efficient FIRNN architecture, several experiments were made varying different parameters. The best predictions were obtained for a network with one input neuron and 10th-order filters in the input layer, two 8-neuron 5th-order filter hidden layers and one output neuron. All the networks were trained with the temporal backpropagation learning algorithm, using the sigmoid transfer function at the hidden layers and a linear output. The learning rate was 0.001. In most experiments a normalized mean square error of 0.4 ± 0.1 and a correlation coefficient between the original and the predicted series greater than 0.8 were found. From a comparison with other SSTA prediction methods, the results obtained with the neural network were the best ones for the short term forecasting case.

LEARNING OF FIR AND ARX NEURAL NETWORKS WITH EMPIRICAL RISK MINIMIZATION ALGORITHM

The probably approximately correct (PAC) learning theory was originally introduced to address static models where the input data were assumed to be i.i.d. In many real applications; however, datasets and systems to be modeled are often dynamic. This encourages the efforts to extend the conventional PAC learning theory to address typical dynamic models such as finite impulse response (FIR) and auto regressive exogenous (ARX) models. This paper presents such extensions for the PAC learning theory and uses the resulting theory to evaluate the learning properties of some families of FIR and ARX neural networks. For ARX models, besides the learning properties of the neural models, stochastic stability of the models are also evaluated.

IDENTIFICATION AND PREDICTION OF UNSTEADY TRANSONIC AERODYNAMIC LOADS BY MULTI-LAYER FUNCTIONALS

Nonlinear unsteady aerodynamic effects present major modelling difficulties in the analysis and control of aeroelastic response. A rigorous mathematical framework, that can account for the complex nonlinearities and time-history effects of the unsteady aerodynamic response, is provided by the use of functional representations. A recent development, based on functional approximation theory, has achieved a new functional form; namely, multi-layer functionals. The development of a multi-layer functional for discrete-time, finite memory, causal systems has been shown to be realizable via finite impulse response neural networks. Identification of an appropriate temporal neural network model of the nonlinear transonic aerodynamic response is facilitated via a supervised training process using multiple input–output sets, with data obtained by an Euler CFD code. The training process is based on a genetic algorithm to optimize the network architecture, combined with a random search algorithm to update weight and bias values. The approach is examined for two different multiple aerodynamic input–output data sets, and in both cases, the prediction properties of the network model establish the multi-layer functional as a suitable representation of unsteady aerodynamic response.

FDTD signal extrapolation using a finite impulse response neural network model

The finite impulse response (FIR) neural network technique is applied as a nonlinear predictor to extrapolate a time-series signal from FDTD simulations. The FIR neural networks are trained by the temporal backpropagation learning algorithm. The model is applied to predict time-domain evolution in the FDTD simulation of a waveguide filter. ©1999 John Wiley & Sons, Inc. Microwave Opt Technol Lett 21: 325–330, 1999.

Modeling judgments on sound quality by neural networks

The current noise control methods are not sufficient for covering the ratings of different kinds of noise. They are mainly based on experiments with stationary and artificial test signals. New judgment models have to be developed to account for the time fluctuations of environmental sounds. During the last decade algorithms emerged in connectionism that not only find optimal mappings within stationary data (for example the backpropagation algorithm) but also within time series. One architecture for this task is called finite impulse response (FIR) neural network. Due to the architecture the networks include preceding time steps for the calculation of the current output. This contribution presents FIR neural networks to model sound quality judgments on environmental sounds. A listening test has been conducted for collecting training samples for the network and for testing its generalization abilities. The sound patterns and the subjects answers are pre-processed using factor analysis. The analysis of the subjects answers yielded 2 factors, which can be identified to represent the sound attributes annoying and powerful. The FIR neural networks are trained to estimate both attributes. The results of training and generalization show high correlation between the sound attributes and predicted judgments of the neural networks.

Access flow control scheme for ATM networks using neural-network-based traffic prediction

The authors propose a new approach to the problem of congestion control arising at the user network interface (UNI) of ATM-based broadband networks. The access flow control mechanism operates on the principle of feedback control. It uses a finite impulse response (FIR) neural network to accurately predict the traffic arrival patterns. The predicted output in conjunction with the current queue information of the buffer can be used as a measure of congestion. When the congestion level is reached, a control signal is generated to throttle the input arrival rate. The FIR multilayer perceptron model and its training algorithm are discussed. Simulation results presented in the paper suggest that the scheme provides a simple and efficient traffic management for ATM networks.

Combination of finite impulse response neural networktechnique with FDTD method for simulation ofelectromagnetic problems

The finite difference time domain (FDTD) method requires long computation times for simulating resonant or high-Q structures. The authors incorporate the finite impulse response neural network technique as a predictor in order to save time in FDTD simulations. The applicability of the technique is demonstrated by carrying out an analysis of a waveguide filter.