※一部利用できない機能があります
| 1.
論文 |
論文
|
Nakayama, Kenji
概要:
金沢大学大学院自然科学研究科情報システム<br />金沢大学工学部<br />A simultaneous frequency- and time-domain approximation method for discrete-time
…
filters is proposed. In the method, transfer function coefficients are divided into two subsets, X//1 and X//2, which are employed for optimizing a time response and a frequency response, respectively. Frequency and time responses are optimized through the iterative Chebyshev approximation method and a method of solving linear equations, respectively. At the rth iteration step, the maximum frequency response error, which appeared at the (r minus 1)th step, is minimized, and X//2**(**r** minus **1**) becomes X//2**(**r**). X//1**(**r**) is obtained from linear equations including X//2**(**r**) as a constant. The frequency response at the rth step is evaluated using the above obtained X//1**(**r**) and X//2**(**r**). This means the optimum time response is always guaranteed in the frequency-response approximation procedure. A design example of a symmetrical impulse response shows the new approach is more efficient than conventional methods from the filter order reduction viewpoint.
続きを見る
|
||||
| 2.
論文 |
論文
|
Hara, Kazuyuki ; Nakayama, Kenji
概要:
金沢大学大学院自然科学研究科知能情報・数理<br />A training data selection method is proposed for multilayer neural networks (MLNNs). This met
…
hod selects a small number of the training data, which guarantee both generalization and fast training of the MLNNs applied to pattern classification. The generalization will be satisfied using the data locate close to the boundary of the pattern classes. However, if these data are only used in the training, convergence is slow. This phenomenon is analyzed in this paper. Therefore, in the proposed method, the MLNN is first trained using some number of the data, which are randomly selected (Step 1). The data, for which the output error is relatively large, are selected. Furthermore, they are paired with the nearest data belong to the different class. The newly selected data are further paired with the nearest data. Finally, pairs of the data, which locate close to the boundary, can be found. Using these pairs of the data, the MLNNs are further trained (Step 2). Since, there are some variations to combine Steps 1 and 2, the proposed method can be applied to both off-line and on-line training. The proposed method can reduce the number of the training data, at the same time, can hasten the training. Usefulness is confirmed through computer simulation.
続きを見る
|
||||
| 3.
論文 |
論文
|
Nakayama, Kenji ; Kato, Takuo ; Katayama, Hiroshi
|
||||
| 4.
論文 |
論文
|
Nakayama, Kenji
概要:
A bandlimited signal extrapolation algorithm is proposed. J. A. Cadzow's algorithm (1979) is modified to eliminate undesired outband spectra. An inverse filter stopband response is relaxed by adding small random numbers. A
…
constrained heuristic optimization is suggested that can use arbitrary signal properties as constraints. Through numerical examples, it is shown how regularization techniques can improve SNR (signal-to-noise-ratio) by 20 dB.
続きを見る
|
||||
| 5.
論文 |
論文
|
Keeni, Kanad ; Shimodaira, Hiroshi ; Nakayama, Kenji
概要:
This paper presents an automatic coding scheme for representing the output layer of a neural network. Compared to local representation where the number of output unit is p, the number of output unit required for the proposed representation is
…
close to log p. The output of seven different printers were used for evaluating the performance of the system. The proposed automatic representation gave the average recognition rate of 98.7% for 71 categories.
続きを見る
|
||||
| 6.
論文 |
論文
|
Wang, Youhua ; Ikeda, Kazushi ; Nakayama, Kenji
|
||||
| 7.
論文 |
論文
|
Nakayama, Kenji ; Katayama, Hiroshi
|
||||
| 8.
論文 |
論文
|
Ikeda, Kazushi ; Suzuki, Akihiro ; Nakayama, Kenji
概要:
The effects of the quantization of the parameters of a learning machine are discussed. The learning coefficient should be as small as possible for a better estimate of parameters. On the other hand, when the parameters are quantized, it
…
should be relatively larger in order to avoid the paralysis of learning originated from the quantization. How to choose the learning coefficient is given in this paper from the statistical point of view.
続きを見る
|
||||
| 9.
論文 |
論文
|
Nakayama, Kenji ; Imai, Kunihiko
概要:
金沢大学理工研究域電子情報学系<br />A neural demodulator is proposed for amplitude shift keying (ASK) signal. It has several important
…
features compared with conventional linear methods. First, necessary functions for ASK demodulation, including wide-band noise rejection, pule waveform shaping, and decoding, can be embodied in a single neural network. This means these functions are not separately designed but unified in a learning and organizing process. Second, these functions can be self-organized through the learning. Supervised learning algorithms, such as the backpropagation algorithm, can be applied for this purpose. Finally, both wide-band noise rejection and a very sharp waveform response can be simultaneously achieved. It is very difficult to be done by linear filtering. Computer simulation demonstrates efficiency of the proposed method.
続きを見る
|
||||
| 10.
論文 |
論文
|
Miyoshi, Seiji ; Nakayama, Kenji
概要:
In this paper, the geometric learning algorithm (GLA) is proposed for an elementary perceptron which includes a single output neuron. The GLA is a modified version of the affine projection algorithm (APA) for adaptive filters.
…
The weights update vector is determined geometrically towards the intersection of the k hyperplanes which are perpendicular to patterns to be classified. k is the order of the GLA. In the case of the APA, the target of the coefficients update is a single point which corresponds to the best identification of the unknown system. On the other hand, in the case of the GLA, the target of the weights update is an area, in which all the given patterns are classified correctly. Thus, their convergence conditions are different. In this paper, the convergence condition of the 1st order GLA for 2 patterns is theoretically derived. The new concept `the angle of the solution area' is introduced. The computer simulation results support that this new concept is a good estimation of the convergence properties.
続きを見る
|
||||
| 11.
論文 |
論文
|
Ma, Zhiqiang ; Nakayama, Kenji ; Yamamoto, G.
|
||||
| 12.
論文 |
論文
|
Nakayama, Kenji ; Kuriki, Satoshi
概要:
A simplified digital tuned-circuit extractor is proposed. An A/D converter and multipliers are not required. An A/D converter and multipliers are not required. The tuning frequency is determined only by a master clock. The tank circuit
…
output noise is sufficiently suppressed by newly introduced error canceller. Although sampling frequency for the tank circuit is relatively low, the phase of the output signal is effectively detected. A computer simulation shows that phase adjusting for data sampling clock is stable. The proposed timing extractor is easily realized on digital integrated circuits.
続きを見る
|
||||
| 13.
論文 |
論文
|
Nakayama, Kenji
概要:
A new difference coefficient FIR digital filter is proposed. Its coefficient is obtained as the difference between the successive values of the original coefficients permuted in large magnitude sequence. It is effectively applied to
…
every kind of filter responses. Computation complexity becomes about 18% and 13% for 99th and 299th order FIR filters, compared with direct FIR filter realization.
続きを見る
|
||||
| 14.
論文 |
論文
|
Miyoshi, Seiji ; Nakayama, Kenji
概要:
金沢大学理工研究域電子情報学系<br />In this paper, probabilistic memory capacity of recurrent neural networks(RNNs) is investigated. Th
…
is probabilistic capacity is determined uniquely if the network architecture and the number of patterns to be memorized are fixed. It is independent from a learning method and the network dynamics. It provides the upper bound of the memory capacity by any learning algorithms in memorizing random patterns. It is assumed that the network consists of N units, which take two states. Thus, the total number of patterns is the Nth power of 2. The probabilities are obtained by discriminations whether the connection weights, which can store random M patterns at equilibrium states, exist or not. A theoretical way for this purpose is derived, and actual calculation is executed by the Monte Carlo method. The probabilistic memory capacity is very important in applying the RNNs to real fields, and in evaluating goodness of learning algorithms. As an example of a learning algorithm, the improved error correction learning is investigated, and its convergence probabilities are compared with the upper bound. A linear programming method can be effectively applied to this numerical analysis.
続きを見る
|
||||
| 15.
論文 |
論文
|
Nakayama, Kenji ; Ikehara, Keisuke
|
||||
| 16.
論文 |
論文
|
Nakayama, Kenji ; Kimura, Yoshinori ; Katayama, Hiroshi
概要:
In this paper, quantization level increase in human face images using a multilayer neural network (NN) is investigated.
…
Basically speaking, it is impossible to increase quality without any other information. However, when images are limited to some category, image restoration could be possible, based on the common properties in this category. The multilayer NN is trained using human face images of 32ﾃ・2 pixels with 8-levels as the input data, and 256-level images as the targets. The standard back-propagation (BP) algorithm is employed. 20, 40 and 100 training data are examined. By increasing the training data, a general function of regenerating missing information can be achieved. The internal structure of the trained NN is analyzed using some special input images. As a result, it has been confirmed that the NN regards the input image as the human face, and extracts features of the face. The input image is transformed using these features and the common properties of the training data, extracted and held on the connection weights, to the human face image.
続きを見る
|
||||
| 17.
論文 |
論文
|
Kuraishi, Yoshiaki ; Makabe, Takayoshi ; Nakayama, Kenji
概要:
金沢大学大学院自然科学研究科情報システム<br />金沢大学工学部<br />This paper presents a fully integrated analog front-end LSI chip which is an inte
…
rface system between digital signal processors and existing analog telecommunication networks. The developed analog LSI chip includes many high level function blocks such as A/D and D/A converters with 11 bit resolution, various kinds of SCF's, an agc circuit, an external control level adjuster, a carrier detector, and a zero crossing detector. Design techniques employed are mainly directed toward circuit size reductions. The LSI chip is fabricated in a 5 mu m line double polysilicon gate NMOS process. Chip size is 7. 14 multiplied by 6. 51 mm. The circuit operates on plus or minus 5 v power supplies. Typical power consumption is 270 mw. By using this analog front-end LSI chip and a digital signal processor, modern systems can be successfully constructed in a compact size.
続きを見る
|
||||
| 18.
論文 |
論文
|
Xu, Q. ; Nakayama, Kenji
概要:
This paper investigates some possible problems of Cascade Correlation algorithm, one of which is the zigzag output mapping caused by weight-illgrowth of the adding hidden unit. Without doubt, it could lead to deteriorate the
…
generalization, especially for regression problems. To solve this problem, we combine Cascade Correlation algorithm with regularization theory. In addition, some new regularization terms are proposed in light of special cascade structure. Simulation has shown that regularization indeed smooth the zigzag out-put, so that the generalization is improved, especially for functional approximation.
続きを見る
|
||||
| 19.
論文 |
論文
|
Nakayama, Kenji ; Chigawa, Yasuhide ; Hasegawa, Osamu
|
||||
| 20.
論文 |
論文
|
20. A weighted competitive learning method extracting skeleton pattern for Japanese Kanji characters
Nakayama, Kenji ; Kato, Takuo
|
||||
| 21.
論文 |
論文
|
Yukawa, Akira ; Maruta, Rikio ; Nakayama, Kenji
|
||||
| 22.
論文 |
論文
|
Nakayama, Kenji ; Iwata, Atsushi ; Yanagisawa, Takeshi
概要:
金沢大学大学院自然科学研究科情報システム
|
||||
| 23.
論文 |
論文
|
Hibino, Masao ; Nakayama, Kenji ; Mizukami, Toshihiko
|
||||
| 24.
論文 |
論文
|
Ohnishi, K. ; Nakayama, Kenji
概要:
金沢大学理工研究域電子情報学系<br />A neural demodulator is proposed for quadrature amplitude modulation (QAM) signals. It has several
…
important features compared with conventional linear methods. First, necessary functions for the QAM demodulation, including wide-band noise rejection, pulse waveform shaping, and decoding, can be embedded in a single neural network. This means that these functions are not separately designed but are unified in a learning process. Second, these functions can be self-organized through the learning. Supervised learning algorithms, such as the back-propagation algorithm, can be applied for this purpose. Finally, both wide-band noise rejection and a very sharp waveform response can be simultaneously achieved. It is very difficult to be done by linear filtering. Computer simulation demonstrates efficiency of the proposed method.
続きを見る
|
||||
| 25.
論文 |
論文
|
Hara, Kazuyuki ; Nakayama, Kenji
概要:
金沢大学理工研究域電子情報学系<br />A training data reduction method for a multilayer neural network (MLNN) is proposed in this paper.
…
This method reduce the data by selecting the minimum number of training data that guarantee generality of the MLNN. For this purpose, two methods are used. One of them is a pairing method which selects the training data by finding the nearest data of the different classes. Data along the class boundary in data space can be selected. The other method is a training method, which used a semi-optimum MLNN in a training process. Since the MLNN classify data based on the distance from the network boundary, the selected data can locate close to the class boundary. So, if the semi-optimum MLNN did not select data from class boundary, pairing method can select them. The proposed methods can be applied to both off-line training and on-line training. The proposed method is also investigated through computer simulation.
続きを見る
|
||||
| 26.
論文 |
論文
|
Hara, Kazuyuki ; Nakayama, Kenji
概要:
Signal classification performance using multilayer neural network (MLNN) and the conventional signal processing methods are theoretically compared under the limited observation period and computational load. The signals with N samples
…
are classified based on frequency components. The comparison is carried out based on degree of freedom the signal detection regions in an N-dimensional signal space. As a result, the MLNN has higher degree of freedom, and can provide more flexible performance for classifying the signals than the conventional methods. This analysis is further investigated throught computer simulations. Multi-frequency signals and the real application, a dial tone receiver, are taken into account. As a result, the MLNN can provide much higher accuracy than the conventional signal processing methods.
続きを見る
|
||||
| 27.
論文 |
論文
|
Nakayama, Kenji ; Sato, Yayoi ; Kuraishi, Yoshiaki
概要:
金沢大学大学院自然科学研究科情報システム<br />金沢大学工学部<br />Design techniques are described for a switched capacitor adaptive line equalizer
…
which is applied to high-speed (200 kb/s) digital transmission over analog subscriber loops. An equalizer transfer function is approximated so as to minimize intersymbol interference of an isolated pulse response. Optimum pole-zero location, which is suited to line characteristics in a wide frequency band, is also discussed. In order to attain high accuracy capacitor ratios using a small unit capacitor, capacitor values are rounded off into equivalent integer values, and are discretely optimized using pole-zero deviation as an error criterion. The equalizer has a finite number of frequency responses which correspond to line lengths. Gain and delay time differences between the adjoining step responses are compressed. The switched capacitor line equalizer was fabricated using 3- mu m CMOS technology. Measured data were very close to designed performances.
続きを見る
|
||||
| 28.
論文 |
論文
|
Nakayama, Kenji ; Mizukami, Toshihiko
概要:
金沢大学大学院自然科学研究科情報システム<br />金沢大学工学部<br />A new infinite impulse response (IIR) Nyquist filter with zero intersymbol interf
…
erence is proposed. The necessary and sufficient conditions for the transfer function are obtained. The proposed IIR Nyquist filter requires only frequency-domain optimization. Multistep optimization, using the iterative Chebyshev approximation, is proposed. This method is able to design a new kind of IIR Nyquist filter with the minimum order. Numerical examples for 30- and 15-percent rolloff rates are illustrated. From these examples, it is confirmed that the IIR approach can reduce the filter order and hardware size, compared with the conventional finite impulse response (FIR) Nyquist filters. Its efficiency becomes marked for high Q Nyquist filters.
続きを見る
|
||||
| 29.
論文 |
論文
|
Nakayama, Kenji
概要:
金沢大学大学院自然科学研究科情報システム<br />This paper proposes a new discrete optimization method which is mainly directed toward saving
…
computing time for high-order FIR filters. A transfer functions first approximated in a cascade form of a low-order function W(z) with pre-rounded coefficients and a high-order function F(z) with infinite precision coefficients. Rounded F(z) coefficients are discretely optimized to minimize the mean square error of the amplitude response. The roundoff error spectrum is shaped to be suppressed by a weighting function W(z). In order to save computing time, the error is equivalently evaluated in a time domain, and the F(z) coefficients are divided into small groups in the discrete optimization procedure. Design examples for 200 tap FIR filters demonstrate practical usefulness.
続きを見る
|
||||
| 30.
論文 |
論文
|
Miyoshi, Seiji ; Ikeda, Kazushi ; Nakayama, Kenji
|
||||
| 31.
論文 |
論文
|
Nakayama, Kenji
|
||||
| 32.
論文 |
論文
|
Hirano, Akihiro ; Nakayama, Kenji ; Watanabe, Kazunobu
概要:
This paper presents convergence characteristics of stereophonic echo cancellers with pre-processing. The convergence analysis of the averaged tap-weights show that the convergence characteristics depends on the relation between the impulse
…
response in the far-end room and the changes of the pre-processing filters. Examining the uniqueness of the solution in the frequency domain leads us to the same relation. Computer simulation results show the validity of these analyses.
続きを見る
|
||||
| 33.
論文 |
論文
|
Nakayama, Kenji ; Ohsugi, M.
|
||||
| 34.
論文 |
論文
|
Nakayama, Kenji ; Mitsutani, N.
概要:
Summary form only given. A novel adaptive threshold method is proposed for an associative memory using a mutually connec
…
ted neural network. Computer simulation was carried out, using 51 and 153 patterns, which appear on a keyboard. The neural network has 16 × 16 = 256 units and full connections. Results demonstrate that dramatic improvements in memory capacity and association rates can be achieved. For example, an association rate for 51 patterns with 40 noises has been increased from 12.2% to 97.7%, compared with a single threshold method.
続きを見る
|
||||
| 35.
論文 |
論文
|
Nakayama, Kenji ; Mizukami, T.
概要:
A new approach for low-bit-rate coding is proposed. Numerical evaluation for the bit-rate compression ratio and SNR shows the proposed method provides good performance compared with conventional linear PCM coding. Furthermore, since the
…
predictor in the coder does not utilize any of the input signal statistics, this method can be applied for both speech and music coding.
続きを見る
|
||||
| 36.
論文 |
論文
|
Hara, Kazuyuki ; Nakayama, Kenji
概要:
Frequency analysis capability of multilayer neural networks, trained by back-propagation (BP) algorithm is investigated. Multi-frequency signal classification is taken into account for this purpose. The number of frequency
…
sets, that is signal groups, is 2approx.5, and the number of frequencies included in a signal group is 3approx.5. The frequencies are alternately located among the signal groups. Through computer simulation, it has been confirmed that the neural network has very high resolution. Classification rates are about 99.5% for training signals, and 99.0% for untraining signals. The results are compared with conventional methods, including Euclidean distance with accuracy of about 65%, Fourier transform with accuracy of about 10approx.30%, and using very high-Q filters with a huge number of computations. The neural network requires only the same number of inner products as the hidden units. Frequency sensitivity and robustness for the random noise are studied. The networks show high frequency sensitivity, namely, the networks have high frequency resolution. Random noise are added to the multi-frequency signals to investigate how does the network cancel uncorrelated noise among the signals. By increasing the number of samples, or training signals, effects of random noise can be cancelled.
続きを見る
|
||||
| 37.
論文 |
論文
|
Khalaf, Ashraf A.M. ; Nakayama, Kenji
|
||||
| 38.
論文 |
論文
|
Nakayama, Kenji ; Takahashi, Yutaka ; Sato, Yayoi ; Nukada, Yasuaki
概要:
金沢大学大学院自然科学研究科情報システム<br />金沢大学工学部<br />An adaptive switched-capacitor (SC) line equalizer system that can be applied to
…
four-wire full-duplex and multirate digital transmission is described. Several kinds of noises can exist in programmable high-gain SC equalizers, such as switching noise, DC offset jump, and transient response. To avoid their effects, an adaptive SC filter is proposed. Two identical SC circuits are used in parallel. An input signal is fed into both SC circuits, and output signals from both circuits are alternately sent after the above noises are eliminated. Furthermore, many different data rates can be handled by slightly modifying an equalizer circuit. A bridged-tap echo canceller and a DC offset canceller are modified so as to be applied to the proposed adaptive SC filter. A line equalizer system, that handles data rates ranging from 3.2 to 64 kb/s was designed. An LSI was fabricated using a 3-μm CMOS process. Experimental results show that noise effects are mostly eliminated, and frequency responses and eye openings are close to the designed values.
続きを見る
|
||||
| 39.
論文 |
論文
|
Khalaf, Ashraf A.M. ; Nakayama, Kenji
概要:
金沢大学大学院自然科学研究科情報システム<br />Time series prediction is very important technology in a wide variety of fields. The actual ti
…
me series contains both linear and nonlinear properties. The amplitude of the time series to be predicted is usually continuous value. For these reasons, we combine nonlinear and linear predictors in a cascade form. The nonlinear prediction problem is reduced to a pattern classification. A set of the past samples x(n - 1), . . . , x(n - N) is transformed into the output, which is the prediction of the next coming sample x(n). So, we employ a multi-layer neural network with a sigmoidal hidden layer and a single linear output neuron for the nonlinear prediction. It is called a Nonlinear Sub-Predictor (NSP). The NSP is trained by the supervised learning algorithm using the sample x(n) as a target. However, it is rather difficult to generate the continuous amplitude and to predict linear property. So, we employ a linear predictor after the NSP. An FIR filter is used for this purpose, which is called a Linear Sub-Predictor (LSP). The LSP is trained by the supervised learning algorithm using also i(n) as a target. In order to estimate the minimum size of the proposed predictor, we analyze the nonlinearity of the time series of interest. The prediction is equal to mapping a set of past samples to the next coming sample. The multi-layer neural network is good for this kind of pattern mapping. Still, difficult mappings may exist when several sets of very similar patterns are mapped onto very different samples. The degree of difficulty of the mapping is closely related to the nonlinearity. The necessary number of the past samples used for prediction is determined by this nonlinearity. The difficult mapping requires a large number of the past samples. Computer simulations using the sunspot data and the artificially generated discrete amplitude data have demonstrated the efficiency of the proposed predictor and the nonlinearity analysis.
続きを見る
|
||||
| 40.
論文 |
論文
|
Miyoshi, Seiji ; Nakayama, Kenji
概要:
A recurrent neural network (RNN), in which each unit has serial delay elements, is proposed for memorizing limit cycles (LCs). This network is called DRNN in this paper. An LC consists of several basic patterns. The hysteresis information
…
of LCs, realized on the connections from the delay elements to the units, is very efficient in the following reasons. First, the same basic patterns can be shared by different LCs. This make it possible to drastically increase the number of LCs, even though using a small number of the basic patterns. Second, noise performance, that is, probability of recalling the exact LC starting from the noisy LC, can be improved. The hysteresis information consists of two components, the order of the basic patterns included in an LC, and the cross-correlation among all the basic patterns. The former is highly dependent on the number of LCs, and the latter the number of all the basic patterns. In order to achieve good noise performance, a small number of the basic patterns is preferred. These properties of the DRNN are theoretically analyzed and confirmed through computer simulations. It is also confirmed that the DRNN is superior to the RNN without delay elements for memorizing LCs.
続きを見る
|
||||
| 41.
論文 |
論文
|
Ma, Zhiqiang ; Ikeda, Tetsuya ; Nakayama, Kenji
|
||||
| 42.
論文 |
論文
|
Nakayama, Kenji ; Inomata, Satoru ; Takeuchi, Yukou
概要:
金沢大学大学院自然科学研究科情報システム<br />A design methodology is presented for digital multilayer neural networks with limited binary e
…
xpressions. An error back-propagation algorithm is modified as follows: the numbers of binary bits used for connections and unit outputs are decreased step by step in the training process. In order to express unit outputs with two-level values, the differential of the logistic function is replaced by a small positive constant used in weight change equations. After the training is completed, binary expressions for connections and unit outputs can be reduced to several-bit and two-level values, respectively. Therefore, no multipliers or nonlinear functions are required in the resulting network, which will be used for pattern recognition. Furthermore, memory capacity and adder circuit hardware can be reduced. The network performance is also insensitive to noisy patterns.
続きを見る
|
||||
| 43.
論文 |
論文
|
Kobori, Hideki ; Ikeda, Kazushi ; Nakayama, Kenji
概要:
金沢大学理工研究域電子情報学系<br />A model of dynamic associative memories is proposed in this paper. The aim is to find all stored pa
…
tterns, and to distinguish the stored and the spurious patterns. Aihara used chaotic neurons and showed that his model has a nonperiodic associative dynamics. In his model, however, it is difficult to distinguish the stored patterns from the others, because the state of the network changes continually. We propose such a new model of neurons that each neuron changes its output to the other when the accumulation of its internal state exceeds a certain threshold. By computer experiments, we show that the state of the network stays at the stored pattern for a while and then travels around to another pattern, and so on. Furthermore, when the number of the stored patterns is small, the stored and the spurious patterns can be distinguished using interval of the network staying these patterns.
続きを見る
|
||||
| 44.
論文 |
論文
|
Nakayama, Kenji
概要:
金沢大学金沢大学大学院自然科学研究科情報システム<br />金沢大学工学部<br />An improved FFT (fast Fourier transform) algorithm combining both decimations
…
in frequency and in time is presented. Stress is placed on a derivation of general formulas for submatrices and multiplicands. Computational efficiency is briefly discussed.
続きを見る
|
||||
| 45.
論文 |
論文
|
Nakayama, Kenji
概要:
金沢大学大学院自然科学研究科情報システム<br />金沢大学工学部<br />A new approach to the implmentation of sum of products (SP) and transposed sum of
…
products (TSP) using partitioned arithmetic is proposed. This aproach makes it possible to realize SP and TSP with recent advanced semiconductor memory technology without multipliers. It is also shown to offer signficant reductions in memory capacity requirement and computational complexity. Dynamic range constraints and output roundoff noise analysis are also discussed.
続きを見る
|
||||
| 46.
論文 |
論文
|
Nakayama, Kenji ; Nishimura, Katsuaki ; Katayama, Hiroshi
概要:
Authors have proposed an asymmetrical associative neural network (NN) using variable hysteresis threshold and its learning and association algorithms. It can drastically improve noise performance, that is insensitivity to noise. In
…
this paper, memory capacity bound and threshold optimization in this associative NN are further discussed. Binary random patterns are considered. First, relation between the number of patterns and the number of iterations is investigated. The latter gradually increases until some number of patterns. After that, it suddenly increases. This is a very peculiar phenomenon. This turning point gives the memory capacity bound, that is about 1.56N, where N is the number of units. Next, threshold optimization is discussed. Relation between threshold and noise performance, and effects of connection weight distribution on noise performance are theoretically discussed. Based on these results, a ratio of step-size and the threshold is optimized to be 0.5/(NP-1), where NP is the number of units on the pattern. Numerically statistical simulation demonstrates efficiency of the proposed methods.
続きを見る
|
||||
| 47.
論文 |
論文
|
Khalaf, Ashraf A.M. ; Nakayama, Kenji
概要:
Time series prediction is a very important technology in a wide variety of field. The actual time series contains both linear and nonlinear properties. The amplitude of the time series to be predicted is usually continuous value. For
…
this reason, we combine nonlinear and linear predictors in a cascade form. In order to estimate the minimum size of the proposed predictor, we propose a nonlinearity analysis for the time series of interest. Computer simulations using the sunspot data have demonstrated the efficiency of the proposed predictor and the nonlinearity analysis.
続きを見る
|
||||
| 48.
論文 |
論文
|
Nakayama, Kenji ; Ma, Zhiqiang ; Yamamoto, G.
概要:
A new pole-zero pairing and ordering strategy which can provide the minimum total capacitance in cascaded biquad switched-capacitor (SC) filters is proposed. Design examples for several kinds of SC filters are demonstrated. The
…
proposed conditions can limit pairing and ordering assignments, by which the total capacitance is concentrated within 5% over the minimum value.
続きを見る
|
||||
| 49.
論文 |
論文
|
Ma, Zhiqiang ; Nakayama, Kenji ; Sugiyama, Akihiko
概要:
金沢大学大学院自然科学研究科情報システム<br />An automatic tap assignment method in subband adaptive filter is proposed in this letter. The
…
number of taps of the adaptive filter in each band is controlled by the mean-squared error. The numbers of taps increase in the bands which have large errors, while they decrease in the bands having small errors, until residual errors in all the bands become the same. In this way, the number of taps in a band is roughly proportional to the length of the impulse response of the unknown system in this band. The convergence rate and the residual error are improved, in comparison with existing uniform tap assignment. Effectiveness of the proposed method has been confirmed through computer simulation.
続きを見る
|
||||
| 50.
論文 |
論文
|
Nakayama, Kenji ; Seki, T.
|
||||
| 51.
論文 |
論文
|
Nakayama, Kenji ; Takahashi, Yutaka ; Satoh, Yayoi ; Naka, Masahiro ; Nukada, Yasuaki
概要:
The authors present a single-chip switched capacitor (SC) line equalizer system which can be applied to full duplex digital transmission. Adaptive SC filters, constructed with programmable capacitor arrays (PCAs), usually cause undesired
…
responses, such as spike noise and transient response, which degrade data transmission quality. To avoid these phenomena, a duplex SC equalizer which has the same circuits in parallel is introduced. The PCAs in one duplex equalizer, whose output is not transferred, are varied. After the undesired phenomena vanish, the output is alternated. The equalizer system can be applied at several different bit rates by changing the external control signals. Algorithms and circuit realizations for a bridged tap echo canceller and a dc offset canceller are further improved.
続きを見る
|
||||
| 52.
論文 |
論文
|
Hara, Kazuyuki ; Nakayama, Kenji
|
||||
| 53.
論文 |
論文
|
Keeni, Kanad ; Nakayama, Kenji ; Shimodaira, Hiroshi
概要:
A method has been proposed for weight initialization in back-propagation feed-forward networks. Training data is analyzed and the notion of critical point is introduced for determining the initial weights and the number of hidden
…
units. The proposed method has been applied to artificial data and the publicly available cancer database. The experimental results of artificial data show that the proposed method takes 1/3 of the training time required for standard back-propagation. In order to verify the effectiveness of the proposed method, standard back-propagation, where the learning starts with random initial weights was also applied to the cancer database. The experimental results indicate that the proposed weight initialization method results in better generalization.
続きを見る
|
||||
| 54.
論文 |
論文
|
Khalaf, Ashraf A.M. ; Nakayama, Kenji
概要:
金沢大学大学院自然科学研究科情報システム<br />A nonlinear time series predictor was proposed, in which a nonlinear sub-predictor (NSP) and a
…
linear subpredictor (LSP) are combined in a cascade form. This model is called hybrid predictor here. The nonlinearity analysis method of the input time series was also proposed to estimate the network size. We have considered the nonlinear prediction problem as a pattern mapping one. A multi-layer neural network, which consists of sigmoidal hidden neurons and a single linear output neuron, has been employed as a nonlinear sub-predictor. Since the NSP includes nonlinear functions, it can predict the nonlinearity of the input time series. However, the prediction is not complete in some cases. Therefore, the NSP prediction error is further compensated for by employing a linear sub-predictor after the NSP. In this paper, the prediction mechanism and a role of the NSP and the LSP are theoretically and experimentally analyzed. The role of the NSP is to predict the nonlinear and some part of the linear property of the time series. The LSP works to predict the NSP prediction error. Furthermore, predictability of the hybrid predictor for noisy time series is investigated. The sigmoidal functions used in the NSP can suppress the noise effects by using their saturation regions. Computer simulations, using several kinds of nonlinear time series and other conventional predictor models, are demonstrated. The theoretical analysis of the predictor mechanism is confirmed through these simulations. Furthermore, predictability is improved by slightly expanding or shifting the input potential of the hidden neurons toward the saturation regions in the learning process.
続きを見る
|
||||
| 55.
論文 |
論文
|
Nakayama, Kenji
概要:
金沢大学大学院自然科学研究科情報システム<br />金沢大学工学部<br />A new approach is proposed for computing the discrete Fourier transform (DFT) wit
…
h power-of-2 length using the butterfly-structure number-theoretic transform (NTT). An algorithm breaking down the DFT matrix into circular matrices with power-of-2 size is introduced. Fast circular convolution, which is implemented by the NTT based on the butterfly structure, provides significant reductions in the number of computations, as well as a simple and regular structure. The proposed algorithm can be successively implemented following a simple flowchart using the reduced-size submatrices. Multiplicative complexity is reduced to about 21% of that with the classical FFT algorithm, preserving almost the same number of additions.
続きを見る
|
||||
| 56.
論文 |
論文
|
Wang, Youhua ; Ikeda, Kazushi ; Nakayama, Kenji
概要:
金沢大学大学院自然科学研究科情報システム<br />金沢大学工学部<br />The numerical property of an adaptive filter algorithm is the most important prob
…
lem in practical applications. Most fast adaptive filter algorithms have the numerical instability problem and the fast Newton transversal filter (FNTF) algorithms are no exception. In this paper, we propose a numerically stable fast Newton type adaptive filter algorithm. Two problems are dealt with in the paper. First, we derive the proposed algorithm from the order-update fast least squares (FLS) algorithm. This derivation is direct and simple to understand. Second, we give the stability analysis using a linear time-variant state-space method. The transition matrix of the proposed algorithm is given. The eigenvalues of the ensemble average of the transition matrix are shown to be asymptotically all less than unity. This results in a much improved numerical performance compared with the FNTF algorithms. The computer simulations implemented by using a finite-precision arithmetic have confirmed the validity of our analysis.
続きを見る
|
||||
| 57.
論文 |
論文
|
Kuraishi, Yoshiaki ; Nakayama, Kenji ; Miyadera, Kazuyuki ; Okamura, Toshiyuki
概要:
金沢大学大学院自然科学研究科情報システム<br />金沢大学工学部<br />A single-chip speech spectrum analyzer which contains a 20-channel filter bank, a
…
9 bit resolution analog-to-digital converter, and a 396 bit buffer memory is described. Several efficient design techniques were used to realize the equivalent 308th-order transfer functions on a single chip monolithic MOS circuit. A new time division multiplexed switched-capacitor filter technique is introduced which can easily cancel dc offsets which appear in the multiplexed channel outputs. The LSI was fabricated in 3. 5 mu m CMOS technology, with a 7. 0 multiplied by 6. 5 mm**2 die size and a 150 mw power consumption with a plus 5 v single power supply. Experimental results show that designed performance was realized.
続きを見る
|
||||
| 58.
論文 |
論文
|
Nakayama, Kenji ; Kimura, Yoshinori
概要:
金沢大学理工研究域電子情報学系<br />An optimization method of activation functions is proposed. Three typical functions are combined in
…
hidden layers. Contribution of the functions is evaluated using three criteria. The useful functions are selected or multiplied in the learning process. Problems of parity and of counting '1' in bit-patterns can be solved by the proposed method with the suitable functions and the minimum number of hidden units.
続きを見る
|
||||
| 59.
論文 |
論文
|
Tomikawa, Y. ; Nakayama, Kenji
概要:
In this paper, a recurrent neural network (RNN) is applied to approximating one to N many valued mappings. The RNN described in this paper has a feedback loop from an output to an input in addition to the conventional multi layer neural
…
network (MLNN). The feedback loop causes dynamic output properties. The convergence property in these properties can be used for this approximating problem. In order to avoid conflict by the overlapped target data y*s to the same input x., the input data set (x*, y*) and the target data y* are presented to the network in learning phase. By this learning, the network function f(x, z) which satisfies y* = f(x*,y*) is formed. In recalling phase, the solutions y of y = f(x,y) are detected by the feedback dynamics of RNN. The different solutions for the same input x can be gained by changing the initial output value of y. It have been presented in our previous paper that the RNN can approximate many valued continuous mappings by introducing the differential condition to learning. However, if the mapping has discontinuity or changes of value number, it sometimes shows undesirable behavior. In this paper, the integral condition is proposed in order to prevent spurious convergence and to spread the attractive regions to the approximating points.
続きを見る
|
||||
| 60.
論文 |
論文
|
Wang, Hong ; Ma, Zhiqiang ; Nakayama, Kenji
概要:
In this paper we investigate the evolutionary heuristics used as approximation algorithm to the subset sum problem. We p
…
ropose a graded penalty function in a fitness function of genetic algorithms to penalize an infeasible string in solving the subset sum problem. An exponential term of generation variable, tθ, is added into the penalty function for increasing penalty generation by generation. The experiments show that the proposed penalty function is more efficient than other existing penalty functions. It is suggested that the penalty pressure is increased step by step.
続きを見る
|
||||
| 61.
論文 |
論文
|
Wang, Youhua ; Nakayama, Kenji
概要:
金沢大学大学院自然科学研究科情報システム<br />The numerical properties of the recursive least squares (RLS) algorithm and its fast versions
…
have been extensively studied. However, very few investigations are reported concerning the numerical behavior of the predictor based least squares (PLS) algorithms that provide the same least squares solutions as the RLS algorithm. This paper presents a comparative study on the numerical performances of the RLS and the backward PLS (BPLS) algorithms. Theoretical analysis of three main instability sources reported in the literature, including the over-range of the conversion factor, the loss of symmetry and the loss of positive definiteness of the inverse correlation matrix, has been done under a finite-precision arithmetic. Simulation results have confirmed the validity of our analysis. The results show that three main instability sources encountered in the RLS algorithm do not exist in the BPLS algorithm. Consequently, the BPLS algorithm provides a much more stable and robust numerical performance compared with the RLS algorithm.
続きを見る
|
||||
| 62.
論文 |
論文
|
Nakayama, Kenji ; Ishikawa, Yutaka ; Kuraishi, Yoshiaki
|
||||
| 63.
論文 |
論文
|
Tomikawa, K. ; Nakayama, Kenji
概要:
金沢大学理工研究域電子情報学系<br />Recurrent Neural Network (RNN) can be applied to solve a combinatorial optimization problem. Howeve
…
r, the existence of the local minima in this model prevents its application to real world. In this paper, an analysis method for network dynamics based on eigenvalues and eigenvectors of a connection weight matrix is proposed. In this analysis, the transition of the number of negative eigenvalues and the movement of eigenspaces by increasing diagonal elements, which correspond to feedback loops of RNN, are discussed. From this analysis, it is confirmed that the number of negative eigenvalues decreases and the eigenspaces move toward the ascent side of the energy slope at the center point of the state space of RNN, as increasing diagonal elements. These behaviors of RNN contribute to the improvement for searching a solution of a combinatorial optimization problem. This analysis method is applied to a 2-dimensional example and 5 cities T.S. problems. From this analysis, it is theoretically found that the network performance of detecting the optimal solution can be improved by increasing the values of diagonal elements of connection matrix.
続きを見る
|
||||
| 64.
論文 |
論文
|
Nakayama, Kenji
概要:
金沢大学大学院自然科学研究科情報システム<br />金沢大学工学部<br />A transfer function is constructed in a cascade form, using a low-order error-fre
…
e function and a high-order function. The high-order function is discretely optimized so that its error spectrum is suppressed by the error-free function. In order to save computing time, the error spectrum is equivalently evaluated in a time domain, and the coefficients are divided into small groups in a discrete optimization procedure
続きを見る
|
||||
| 65.
論文 |
論文
|
Nakayama, Kenji ; Hirano, Akihiro ; Sakaguchi, H.
概要:
A sub-band adaptive filter, in which modulation and demodulation are employed, were proposed. The sampling rate can be minimized, while no aliasing occurs. In this paper, a polyphase and FFT realization of this sub-band adaptive filter is
…
proposed. In the polyphase and FFT realization for the transmultiplexer, the sampling rate reduction is the same as the number of the sub-bands. In the proposed method, however, they are different, so the conventional method cannot be applied. In the new realization, the polyphase filters are divided into a tapped delay line and an FFT part so as to be shared by all the sub-bands. Computational complexity is well reduced compared with the direct realization.
続きを見る
|
||||
| 66.
論文 |
論文
|
Yukawa, Akira ; Nakayama, Kenji ; Kawakami, Yuichi ; Hinooka, Kiyonobu ; Mizukami, Yoshihiko
|
||||
| 67.
論文 |
論文
|
Nakayama, Kenji ; Nishimura, Katsuaki
概要:
金沢大学理工研究域電子情報学系<br />An associative memory using fixed and variable hysteresis thresholds in learning and recalling proc
…
esses, respectively, has been proposed by authors. This model can achieve a large memory capacity and very low noise sensitivity. However, a relation between weight change Δ w and the hysteresis threshold ± T has not been well discussed. In this paper, a new learning algorithm is proposed, which is based on a delta rule. However, in order to stabilize the learning process, a method of using double hysteresis thresholds is proposed. Unit states are updated using ± T. The error, used for adjusting weights, is evaluated using ± (T+dT). This means 'over correction'. Stable and fast convergence can be obtained. Relations between η =dT/T and convergence rate and noise sensitivity are discussed, resulting the optimum selection for η. Furthermore, the order of presenting training data is optimized taking correlation, into account. In the recalling process, a threshold control method is further proposed in order to achieve fast recalling from noisy patterns.
続きを見る
|
||||
| 68.
論文 |
論文
|
Minamimoto, Kazuhiro ; Ikeda, Kazushi ; Nakayama, Kenji
概要:
In order to analyze the topological structure of the data space using Kohonen's self-organizing feature map (SOFM), a criterion is discussed. The Euclidian distance between the reference vector and the data, the number of the reference
…
vectors and the topology preserving measure are taken into account, and are combined in a unified criterion. Through computer simulation, it is confirmed that goodness of the different reference topologies, that is dimensions, can be clearly discriminated regardless the parameters. Thus, the unified criterion makes it possible to analyze the essential data space topology.
続きを見る
|
||||
| 69.
論文 |
論文
|
Nakayama, Kenji ; Hirano, Akihiro ; Ido, Issei
概要:
Network size of neural networks is highly dependent on activation functions. A trainable activation function has been proposed, which consists of a linear combination of some basic functions. The activation functions and the connection
…
weights are simultaneously trained. An 8 bit parity problem can be solved by using a single output unit and no hidden unit. In this paper, we expand this model to multilayer neural networks. Furthermore, nonlinear functions are used at the unit inputs in order to realize more flexible transfer functions. The previous activation functions and the new nonlinear functions are also simultaneously trained. More complex pattern classification problems can be solved with a small number of units and fast convergence.
続きを見る
|
||||
| 70.
論文 |
論文
|
Hara, Kazuyuki ; Nakayama, Kenji
概要:
金沢大学大学院自然科学研究科情報システム<br />In this paper, a training data selection method for multilayer neural networks (MLNNs) in on-l
…
ine training is proposed. Purpose of the reduction in training data is reducing the computation complexity of the training and saving the memory to store the data without losing generalization performance. This method uses a pairing method, which selects the nearest neighbor data by finding the nearest data in the different classes. The network is trained by the selected data. Since the selected data located along data class boundary, the trained network can guarantee generalization performance. Efficiency of this method for the on-line training is evaluated by computer simulation.
続きを見る
|
||||
| 71.
論文 |
論文
|
Hernandez, C.A. ; Nakayama, Kenji ; Fernandez, M.
概要:
We present a new extension of the Backpropagation learning algorithm by using interval arithmetic. The proposed algorithm represents a generalization of Backpropagation and contains Backpropagation like a particular case. This new
…
algorithm permits the use of training samples and targets which can be indistinctly points and intervals. Among the possible applications of this algorithm, we report its usefulness to integrate expert's knowledge and experimental samples and also its ability to handle 'don't care attributes' in a simple and natural way in comparison with Backpropagation. It also adds flexibility to the codification of inputs and outputs.
続きを見る
|
||||
| 72.
論文 |
論文
|
Nakayama, Kenji
概要:
金沢大学大学院自然科学研究科情報システム<br />金沢大学工学部<br />new realization for FIR digital filters, using permuted difference coefficients,
…
is proposed. Its coefficients are obtained as the difference between the successive values of the original coefficients reordered in a sequence with falling magnitude. The proposed realization can hold desirable properties in FIR filters, such as an exactly linear phase characteristic and stable implementation, and is effectively applied to a wide range filter response. Quantization error analysis shows that the internal data word lengths must be somewhat increased to maintain the same roundoff noise as in a direct form realization. Computational complexity becomes about 23% and 18% for 99th- and 299th-order filters, taking the excess data word lengths into account, compared with the direct form.
続きを見る
|
||||
| 73.
論文 |
論文
|
Nakayama, Kenji ; Katayama, Hiroshi
概要:
A new low-bit learning algorithm for digital multilayer neural networks applied to pattern recognition is proposed. The training can be carried out with a small number of bits. To make the neural network insensitive to noisy patterns,
…
conditions to be satisfied by hidden layer outputs are discussed. Based on this, optimum targets are assigned to the hidden layers. Computer simulation demonstrated the efficiency of the proposed method.
続きを見る
|
||||
| 74.
論文 |
論文
|
Maruta, Rikio ; Kanemasa, Akira ; Sakaguchi, Hisashi ; Hibino, Masao ; Nakayama, Kenji
概要:
金沢大学大学院自然科学研究科情報システム<br />金沢大学工学部<br />Two new digital transmultiplexers intended for commerical use have been developed
…
. One transmultiplexer performs a bilateral conversion between two 12-channel FDM group signals and a 24-channel PCM carrier signal. The other mutually connects two 60-channel FDM supergroup signals and five 24-channel or four 30-channel PCM signals. Both exploit a block processing digital SSB-FDM multiplex/demultiplex scheme employing a cascade of an FFT processor and a set of complex coefficient digital filters. They have been built using newly developed high-level DSP LSI chips. Algorithmic considerations, developed LSI architecture, and equipment configuration are described as well as digital processor design details and measured performance.
続きを見る
|
||||
| 75.
論文 |
論文
|
Nakayama, Kenji ; Kuriki, Satoshi
概要:
金沢大学理工研究域 電子情報学系
|
||||
| 76.
論文 |
論文
|
Wang, Youhua ; Nakayama, Kenji ; Ma, Zhiqiang
概要:
金沢大学理工研究域 電子情報学系<br />This paper presents a new structure for noise and echo cancelers based on a combined fast adaptive
…
algorithm. The main purpose of the new structure is to detect both the double-talk and the unknown path change. This goal is accomplished by using two adaptive filters. A main adaptive filter Fn, adjusted only in the non-double-talk period by the normalized LMS algorithm, is used for providing the canceler output. An auxiliary adaptive filter Ff, adjusted by the fast RLS algorithm, is used for detecting the double-talk and obtaining a near optimum tap-weight vector for Fn in the initialization period and whenever the unknown path has a sudden or fast change. The proposed structure is examined through computer simulation on a noise cancellation problem. Good cancellation performance and stable operation are obtained when signal is a speech corrupted by a white noise, a colored noise and another speech signal. Simulation results also show that the proposed structure is capable of distinguishing the near-end signal from the noise path change and quickly tracking this change.
続きを見る
|
||||
| 77.
論文 |
論文
|
Yapi, Atse ; Ma, Zhiqiang ; Nakayama, Kenji
|
||||
| 78.
論文 |
論文
|
Keeni, Kanad ; Nakayama, Kenji ; Shimodaira, Hiroshi
概要:
金沢大学理工研究域 電子情報学系
|
||||
| 79.
論文 |
論文
|
Nakayama, Kenji
概要:
金沢大学理工研究域 電子情報学系
|
||||
| 80.
論文 |
論文
|
Wang, Hong ; Ma, Zhiqiang ; Nakayama, Kenji
概要:
金沢大学理工研究域 電子情報学系<br />In this paper we investigate the evolutionary heuristics used as approximation algorithm to the su
…
bset sum problem. We propose a graded penalty function in a fitness function of genetic algorithms to penalize an infeasible string in solving the subset sum problem. An exponential term of generation variable, tθ, is added into the penalty function for increasing penalty generation by generation. The experiments show that the proposed penalty function is more efficient than other existing penalty functions. It is suggested that the penalty pressure is increased step by step.
続きを見る
|
||||
| 81.
論文 |
論文
|
Kanemasa, A. ; Maruta, R. ; Nakayama, Kenji ; Sakamura, Y. ; Tanaka, S.
概要:
金沢大学理工研究域 電子情報学系
|
||||
| 82.
論文 |
論文
|
Wang, Y. ; Nakayama, Kenji
|
||||
| 83.
論文 |
論文
|
Hara, Kazuyuki ; Nakayama, Kenji
概要:
This paper discusses properties of activation functions in multilayer neural network applied to pattern classification. A rule of thumb for selecting activation functions or their combination is proposed. The sigmoid, Gaussian and
…
sinusoidal functions are selected due to their independent and fundamental space division properties. The sigmoid function is not effective for a single hidden unit. On the contrary, the other functions can provide good performance. When several hidden units are employed, the sigmoid function is useful. However, the convergence speed is still slower than the others. The Gaussian function is sensitive to the additive noise, while the others are rather insensitive. As a result, based on convergence rates, the minimum error and noise sensitivity, the sinusoidal function is most useful for both without and with additive noise. Property of each function is discussed based on the internal representation, that is the distributions of the hidden unit inputs and outputs. Although this selection depends on the input signals to be classified, the periodic function can be effectively applied to a wide range of application fields.
続きを見る
|
||||
| 84.
論文 |
論文
|
84. A weighted competitive learning method extracting skeleton pattern for Japanese Kanji characters
Nakayama, Kenji ; Kato, Takuo
概要:
金沢大学理工研究域 電子情報学系<br />A weighted competitive learning (WCL) method was proposed by authors for extracting skeleton patte
…
rns from digit and alphabet characters. The extracted pattern is essential in character recognition. It can satisfy the following important requirements. (a) Insensitive to irregular edge lines. (b) Non-structure patterns are not extracted. (c) Insensitive to non-uniform line width. (d) Line information should be held even though the line width widely changes in a character. In this paper, the previous WCL method is improved for application to more complicated characters, such as Japanese Kanji characters. Furthermore, a PDP model, implements the WCL method, is provided.
続きを見る
|
||||
| 85.
論文 |
論文
|
Nakayama, Kenji ; Imai, Kunihiko
概要:
金沢大学理工研究域 電子情報学系
|
||||
| 86.
論文 |
論文
|
Chen, Y. ; Nakayama, Kenji
概要:
An extrapolation algorithm for discrete bandlimited signals is proposed. The bandlimited condition is satisfied by using an ideal bandlimited signal as an excited signal. A waveform synthesis digital filter is used to extrapolate
…
fractionally observed data, using the bandlimited excited signal. Therefore, the proposed method reduces the extrapolation problem to a digital filter estimation problem. Digital filter coefficients are obtained by solving linear equations, which are formulated to minimize the mean-square error between a part of the extrapolated signal and the fractionally observed data. The modified Gram-Schmidt (MGS) procedure is applied to finding the least-square solution of the filter coefficients, due to its superior numerical stability. The recursive version of the MGS procedure is suited to real-time hardware implementation. The proposed algorithm is shown to have good performance in extrapolating the fractional data, observed in a single as well as multiple time windows. This concept is applicable not only to the signal extrapolation but also to the spectral estimation.
続きを見る
|
||||
| 87.
論文 |
論文
|
Nakayama, Kenji ; Ma, Zhiqiang
概要:
金沢大学理工研究域 電子情報学系
|
||||
| 88.
論文 |
論文
|
Kuranishi, Yoshiaki ; Takahashi, Yutaka ; Nakayama, Kenji ; Senba, Takashi
概要:
金沢大学理工研究域 電子情報学系
|
||||
| 89.
論文 |
論文
|
Hara, Kazuyuki ; Nakayama, Kenji
概要:
金沢大学理工研究域 電子情報学系
|
||||
| 90.
論文 |
論文
|
Hara, Kazuyuki ; Nakayama, Kenji
概要:
金沢大学理工研究域 電子情報学系<br />This paper compares signal classification performance of multilayer neural networks (MLNNs) and li
…
near filters (LFs). The MLNNs are useful for arbitrary waveform signal classification. On the other hand, LFs are useful for the signals, which are specified with frequency components. In this paper, both methods are compared based on frequency selective performance. The signals to be classified contain several frequency components. Furthermore, effects of the number of the signal samples are investigated. In this case, the frequency information may be lost to some extent. This makes the classification problems difficult. From practical viewpoint, computational complexity is also limited to the same level in both methods. IIR and FIR filters are compared. FIR filters with a direct form can save computations, which is independent of the filter order. IIR filters, on the other hand, cannot provide good signal classification due to their phase distortion, and require a large amount of computations due to their recursive structure. When the number of the input samples is strictly limited, the signal vectors are widely distributed in the multi-dimensional signal space. In this case, signal classification by the LF method cannot provide a good performance. Because, they are designed to extract the frequency components. On the other hand, the MLNN method can form class regions in the signal vector space with high degree of freedom. When the number of the signal samples is not so limited, both the MLNN and LF methods can provide the same high classification rates. In this case, since the signal vectors are distributed in the specific region, the MLNN method has some convergence problem, that is local minimum problem. The initial weights should be carefully determined around the optimum solution. Another point is robustness for noisy signal. The LFs can suppress wide-band noise by using very high-Q filters. However, the MLNN method can be also robust. Rather, it is a little superior to the LF method when the computational load is limited.
続きを見る
|
||||
| 91.
論文 |
論文
|
Ikeda, Kazushi ; Wang, Youhua ; Nakayama, Kenji
概要:
金沢大学理工研究域電子情報学系<br />The numerical property of the recursive least squares (RLS) algorithm has been extensively studied.
…
However, very few investigations are reported concerning the numerical behavior of the predictor-based least squares (PLS) algorithms which provide the same least squares solutions as the RLS algorithm. In Ref. [9], we gave a comparative study on the numerical performances of the RLS and the backward PLS (BPLS) algorithms. It was shown that the numerical property of the BPLS algorithm is much superior to that of the RLS algorithm under a finite-precision arithmetic because several main instability sources encountered in the RLS algorithm do not appear in the BPLS algorithm. This paper theoretically shows the stability of the BPLS algorithm by error propagation analysis. Since the time-variant nature of the BPLS algorithm, we prove the stability of the BPLS algorithm by using the method as shown in Ref. [6]. The expectation of the transition matrix in the BPLS algorithm is analyzed and its eigenvalues are shown to have values within the unit circle. Therefore we can say that the BPLS algorithm is numerically stable.
続きを見る
|
||||
| 92.
論文 |
論文
|
Kuranishi, Yoshiaki ; Nakayama, Kenji ; Miyadera, Kazuyuki ; Okumura, Toshiyuki
概要:
金沢大学理工研究域 電子情報学系<br />A single-chip speech spectrum analyzer which contains a 20-channel filter bank, a 9-bit-resolution
…
analog-to-digital converter, and a 396-bit buffer memory is described. Several efficient design techniques were used to realize the equivalent 308th-order transfer functions on a single chip monolithic MOS circuit. A new time-division-multiplexed switched-capacitor filter technique is introduced which can easily cancel DC offsets which appear in the multiplexed channel outputs. The LSI was fabricated in 3.5-μm CMOS technology, with a 7.0×6.5 mm/SUP 2/ die size, a power consumption of 150 mW, with a single power supply of +5 V. Experimental results show that designed performance was realized.
続きを見る
|
||||
| 93.
論文 |
論文
|
Nakayama, Kenji ; Chigawa, Y.
|
||||
| 94.
論文 |
論文
|
Hara, Kazuyuki ; Nakayama, Kenji
概要:
金沢大学理工研究域 電子情報学系
|
||||
| 95.
論文 |
論文
|
Nakayama, Kenji ; Katayama, Hiroshi
|
||||
| 96.
論文 |
論文
|
Nakayama, Kenji ; Hasegawa, H. ; Hernandez, C.A.
概要:
金沢大学理工研究域 電子情報学系
|
||||
| 97.
論文 |
論文
|
Wang, Youhua ; Nakayama, Kenji
概要:
金沢大学理工研究域 電子情報学系<br />This paper proposes a new combined fast algorithm for transversal adaptive filters. The fast trans
…
versal filter (FTF) algorithm and the normalized LMS (NLMS) are combined in the following way. In the initialization period, the FTF is used to obtain fast convergence. After converging, the algorithm is switched to the NLMS algorithm because the FTF cannot be used for a long time due to its numerical instability. Nonstationary environment, that is, time varying unknown system for instance, is classified into three categories: slow time varying, fast time varying and sudden time varying systems. The NLMS algorithm is applied to the first situation. In the latter two cases, however, the NLMS algorithm cannot provide a good performance. So, the FTF algorithm is selected. Switching between the two algorithms is automatically controlled by using the difference of the MSE sequence. If the difference exceeds a threshold, then the FTF is selected. Other wise, the NLMS is selected. Compared with the RLS algorithm, the proposed combined algorithm needs less computation, while maintaining the same performance. Furthermore, compared with the FTF algorithm, it provides numerically stable operation.
続きを見る
|
||||
| 98.
論文 |
論文
|
Nakayama, Kenji ; Uriya, Susumu ; Matsuura, Takashi ; Mitama, Masataka
概要:
金沢大学理工研究域 電子情報学系
|
||||
| 99.
論文 |
論文
|
Hara, Kazuyuki ; Nakayama, Kenji
概要:
金沢大学理工研究域 電子情報学系<br />Signal classification performance using multilayer neural network (MLNN) and the conventional sign
…
al processing methods are theoretically compared under the limited observation period and computational load. The signals with N samples are classified based on frequency components. The comparison is carried out based on degree of freedom the signal detection regions in an N-dimensional signal space. As a result, the MLNN has higher degree of freedom, and can provide more flexible performance for classifying the signals than the conventional methods. This analysis is further investigated throught computer simulations. Multi-frequency signals and the real application, a dial tone receiver, are taken into account. As a result, the MLNN can provide much higher accuracy than the conventional signal processing methods.
続きを見る
|
||||
| 100.
論文 |
論文
|
Huq, Asadul ; Ma, Zhiqiang ; Nakayama, Kenji
|
