The conventional linear prediction can be viewed as a constrained blind equalization problem that has gained a lot of interests along with development of telecommunication networks. Because the blind equalization or deconvolution is a general framework of the inverse problem, the reliable and faster algorithm is requested in many applications. This paper proposes an orthogonal wavelet transform domain realization of a blind equalization technique termed as EVA, and presents an application to speech analysis. An orthogonal transformation has no influence to the equalization result in general, but we show that a particular wavelet makes the matrix in EVA nearly lower triangular that promotes the faster convergence in the estimation of maximum eigenvalue and its associate vector in EVA iteration. The experiments with the Japanese vowels show that the the proposed method effectively separates the glottis and vocal tract information, hence is promising for speech analysis.

We present a new method for estimation of spectrum transition of nonstationary signals in cases of low signal-to-noise ratio (SNR). Instead of the basic functions employed in the previously proposed time-varying autoregressive (AR) modeling, we introduce a spectrum transition constraint into the cost function described by the partial correlation (PARCORR) coefficients so that the method is applicable to noisy nonstationary signals of which spectrum transition patterns are complex. By applying this method to the analysis of vibration signals on the interventricular septum (IVS) of the heart, noninvasively measured by the novel method developed in our laboratory using ultrasonics, the spectrum transition pattern is clearly obtained during one cardiac cycle for normal subjects and a patient with cardiomyopathy.

In this paper, the correlation properties are used to develop two efficient encoding schemes for speech line spectrum frequency (LSF) parameters. The first scheme (1D KL), which exploits the intraframe correlation, is based on one-dimensional Karhunen-Loeve (KL) transformation; the second scheme, which requires some coding delays to further utilize the interframe correlation, uses two-dimensional (2D KL) transform in the frequency domain or one-dimensional KL transform co-operating with DPCM in the time domain. Moreover, since the KL transform is globally optimal, which is sensitive to the change of input data statistics, further two adaptive transform coding systems are also investigated in this paper. The performance of all systems for different bit rates is investigated and adequate comparisons are made. It is shown that the gain of using KL transformation to exploit the intraframe and interframe correlation is 3 and 4 bits/speech frame, respectively.

A new method of designing a microphone array with two outputs preserving binaural information is proposed in this paper. This system employs adaptive beamforming using multiple constraints. The binaural cues may be preserved in the two outputs by use of these multiple constraints with simultaneous beamforming to enhance target signals is also available. A computer simulation was conducted to examine the performance of the beamforming. The results showed that the proposed array can perform both the generation of the binaural cues and the beamforming as intended. In particular, beamforming with double-constraints exhibits the best performance; DI is around 7 dB and good interchannel (interaural) time/phase and level differences are generated within a target region in front. With triple-constraints, however, the performance of the beamforming becomes poorer while the binaural information is better realized. Setting of the desired responses to give proper binaural information seems to become critical as the number of the constraints increases.

Under a contamination of background sound noises, it seems difficult especially in a real working situation to evaluate various type statistics of only an objective sound signal fluctuation. In many cases of the noise evaluation, some signal processing method have been employed to eliminate the effect of background sound noises by first measuring emitted sound levels. In this study, a new evaluation method of sound level fluctuation is proposed in principle on the basis of the measurement of heterogeneous physical quantity other than sound pressures or sound levels to eliminate the effect of background sound noises. Though the theoretical analysis on acoustical emission caused by a mechanical vibration seems very difficult in a working situation, the sound noise fluctuation emitted only from an objective sound source can be effectively evaluated through its related vibration measurement by employing a fairly unified stochastic method proposed on the basis of a generalized regression analysis between sound and vibration. Here, the regression coefficients are determined by employing the least squares error method to minimize the mean square of estimation error to illustrate well the sound data by means of vibration data. Finally, the effectiveness of proposed method has been experimentally applied to the sound noise evaluation of a jigsaw.

In this paper, an adaptive method for active control of vibration intensity in the frequency domain is proposed. In this method, vibration intensity is observed with the 4-sensor method, and the coefficients of an adaptive FIR filter for the active control is renewed with the Block Filtered-X LMS algorithm in the frequency domain. An experiment with the proposed method is performed on a simple model. As a result, the proposed method gives larger attenuation of vibration intensity than the conventional method in the high frequency region. The overall attenuation in vibration intensity in that frequency region is 14.1 dB with the proposed method, while it is 7.0 dB with the conventional method. In the lower frequency region, the reduction in vibration intensity by the proposed method is roughly equivalent to that obtained by the conventional method. An improvement may also be achieved there by setting the intervals between error sensors properly.

This paper proposes an adaptive microphone array using blind deconvolution. The method realizes an signal enhancement based on the combination of blind deconvolution, synchronized summation and DSA (Delay-and-Sum Array) method. The proposed method improves performance of estimation by the iterative operation of blind deconvolution using a cost-function based on the coherency function.

In order that the speech recognition system might have a high performance in the noisy environment, the directional microphone arrays at the input of the system have been broadly investigated. The purpose of this study is to develop a new wide-band directional microphone system in view of advancing to an adaptive one afterwards. In the proposed system, three microphones are arranged on a straight line and the beamforming is accomplished in such a way that the output value of the middle microphone is added to the integrated value of the difference between two microphones at both sides. In this study, the signal processing of microphone outputs is implemented by using active RC circuits. Finally, the objective directivity can be experimentally obtained in wide frequency ranges required for the speech recognition.

This paper proposes a method of improving the accuracy of the attenuation constant estimate obtained by using the cross-spectral technique. In the cross-spectral technique, the envelope of the estimated impulse response is deformed due to the use of a time window. As a result, the estimated impulse response decays more rapidly than the real impulse response does, and the attenuation constant obtained by the estimated impulse response becomes larger than the real value. This paper first describes how the attenuation constant changes in the process of impulse response estimation. Next, we propose a method of improving the accuracy of the estimation. The effect of the proposed method is confirmed by computer simulation.

An Acoustic echo canceller has problems adaptating under noisy or double-talk conditions. The adaptation process requires a precise identification of the temporarily changed room impulse response. To do this, both minimizing the step size parameter of the Least Mean Square (LMS) method to be as small as possible and giving up on updating the adaptive filter coefficients have been considered. This paper describes an adaptive cross-spectral technique that is robust to adaptive filtering under noisy or double-talk conditions and for colored signals such a speech signal. The cross-spectral technique was originally developed to measure the impulse response in a linear system. Here we apply in the adaptive cross-spectral technique to solve the acoustic echo cancelling problem. This cross-spectral technique takes the ensemble average of the cross spectrum between input and error signals and the averaged cross spectrum is divided by the averaged power spectrum of the input signal to update the filter coefficients. We have confirmed that the echo signal is suppressed by about 15 dB even under double-talk conditions. We also explain that this method has a systematic error due to using a short time block for estimating the room impulse response. Then we investigate overlapping every last half block by the following first half block in order to reduce the effect of the systematic error. Finally, we compare our method with the Frequency-domain Block LMS (FBLMS) method because both methods are implemented in the frequency domain using a short time block.

This paper proposes a new robust adaptive beamformer applicable to microphone arrays. The proposed beamformer is a generalized sidelobe canceller (GSC) with a variable blocking matrix using coefficient-constrained adaptive filters (CCAFs). The CCAFs, whose common input signal is the output of a fixed beamformer, minimize leakage of the target signal into the interference path of the GSC. Each coefficient of the CCAFs is constrained to avoid mistracking. In the multiple-input canceller, leaky adaptive filters are used to decrease undesirable target-signal cancellation. The proposed beamformer can allow large look-direction error with almost no degradation in interference-reduction performance and can be implemented with a small number of microphones. The maximum allowable look-direction error can be specified by the user. Simulation results show that the proposed beamformer, when designed to allow about 20of look-direction error, can suppress interference by more than 17 dB.

Enhancement of resonance is shown by coupling and summing in sinusoidally driven chaotic neural networks. This resonance phenomenon has a peak at a drive frequency similar to noise-induced stochastic resonance (SR), however, the mechanism is different from noise-induced SR. We numerically study the properties of resonance in chaotic neural networks in the turbulent phase with summing and homogeneous coupling, with particular consideration of enhancement of the signal-to-noise ratio (SNR) by coupling and summing. Summing networks can enhance the SNR of a mean field based on the law of large numbers. Global coupling can enhance the SNR of a mean field and a neuron in the network. However, enhancement is not guaranteed and depends on the parameters. A combination of coupling and summing enhances the SNR, but summing to provide a mean field is more effective than coupling on a neuron level to promote the SNR. The global coupling network has a negative correlation between the SNR of the mean field and the Kolmogorov-Sinai (KS) entropy, and between the SNR of a neuron in the network and the KS entropy. This negative correlation is similar to the results of the driven single neuron model. The SNR is saturated as an increase in the drive amplitude, and further increases change the state into a nonchaotic one. The SNR is enhanced around a few frequencies and the dependence on frequency is clearer and smoother than the results of the driven single neuron model. Such dependence on the drive amplitude and frequency exhibits similarities to the results of the driven single neuron model. The nearest neighbor coupling network with a periodic or free boundary can also enhance the SNR of a neuron depending on the parameters. The network also has a negative correlation between the SNR of a neuron and the KS entropy whenever the boundary is periodic or free. The network with a free boundary does not have a significant effect on the SNR from both edges of the free boundaries.

A novel method to enhance the practical security of interferometric quantum cryptography is proposed, giving the protocol and detailed constructions including a controlled spontaneous photon emitter, a superradiance amplifier, beam splitters, phase shifters, and a pair of Mach-Zehnder interferometers. The intrinsic uncertainty due to the random phase selection out of three, leads to the detection of eavesdropping. The physical uncertainty of the controlled spontaneous emission of coherent photons also adds temporal equivocation to confuse eavesdroppers.

Recently, Miura introduced a construction method of one-point algebraic geometry codes on singular curves, which is regarded as a generalization of one on nonsingular curves, and enables us to construct codes on wider class of algebraic curves. However, it is still not clear whether there really exist singular curves on which we can construct good codes that are never obtained from nonsingular curves. In this paper, we show that for fixed designed minimum distance in a wide range, the dimension of codes on a singular curve is smaller than or equal to that of the codes on its normalization, and the number of check symbols of the former codes is larger than that of the latter codes. This implies the optimality of nonsingular curves for code construction.

We show by numerical calculations that a chaotic neuron model driven by a weak sinusoid has resonance. This resonance phenomenon has a peak at a drive frequency similar to that of noise-induced stochastic resonance (SR). This neuron model was proposed from biological studies and shows a chaotic response when a parameter is varied. SR is a noise induced effect in driven nonlinear dynamical systems. The basic SR mechanism can be understood through synchronization and resonance in a bistable system driven by a subthreshold sinusoid plus noise. Therefore, background noise can boost a weak signal using SR. This effect is found in biological sensory neurons and obviously has some useful sensory function. The signal-to-noise ratio (SNR) of the driven chaotic neuron model is improved depending on the drive frequency; especially at low frequencies, the SNR is remarkably promoted. The resonance mechanism in the model is different from the noise-induced SR mechanism. This paper considers the mechanism and proposes possible explanations. Also, the meaning of chaos in biological systems based on the resonance phenomenon is considered.

In this paper, we propose a new competitive learning algorithm for training single-layer neural networks to cluster data. The proposed algorithm adopts a new measure based on the idea of "symmetry" so that neurons compete with each other based on the symmetrical distance instead of the Euclidean distance. The detected clusters may be a set of clusters of different geometrical structures. Four data sets are tested to illustrate the effectiveness of our proposed algorithm.

A model of the prefrontal cortical circuit has been constructed to investigate the dynamics for working memory processing. The model circuit is multi-layered and consists of a number of circuit modules or columns, each of which has local, excitatory and inhibitory connections as well as feedback connections. The columns interact with each other via the long-range horizontal connections. Besides these intrinsic connections, the pyramidal and spiny cells in the superficial layers receive the specific cue-related input and all the cortical neurons receive a hypothetical bias input. The model cortical circuit amplifies the response to the transient, cue-related input. The dynamics of the circuit evolves autonomously after the termination of the input. As a result, the circuit reaches in several hundred milliseconds an equilibrium state, in which the neurons exhibit graded-level, sustained activity. The sustained activity varies gradually with the cue direction, thus forming memory fields. In the formation of the memory fields, the feedback connections, the horizontal connections, and the bias input all play important roles. Varying the level of the bias input dramatically changes the dynamics of the model cortical neurons. The computer simulations show that there is an optimum level of the input for the formation of well-defined memory fields during the delay period.

We consider a system of coupled two oscillators with external force. At first we introduce the symmetrical property of the system. When the external force is not applied, the two oscillators are synchronized at the opposite phase. We obtain a bifurcation diagram of periodic solutions in the coupled system when the single oscillator has a stable anti-phase solution. We find that the synchronized oscillations eventually become in-phase when the amplitude of the external force is increased.

An optimun design for N(arbitrary)-sheet capacitive Jaumann elctromagnetic (EM) wave absorber, using genetic algorithm will be presented. This algorithm is a random optimization method based on the genetic relation in the human being. We show the bandwidth for two-sheet capacitive Jaumann absorber can be expanded even more than 108% showed by knott, by using this algorithm and without imposing the double-notch design criteria. We also show that our results approaches knott's results when we restrict the characteristic impedances and lengths of the lines to vary within a very short range. We also design one-sheet and three-sheet capacitive Jaumann absorbers. The only restriction used here is about the meaningful range for the design variables. The goal of this algorithm is that we can impose arbitrary restriction about the range of the variation of the variables. So we can see the performance behaviour with the range dimension of the variables, and we can obtain different optimum results for different ranges. Finally we obtain a 20-dB attenuation bandwidth more than 145% for one-sheet, 173% for two-sheet (compare with 108% obtained in [1]) and 193% for three-sheet capacitive Jaumann EM absorbers, with some acceptable short range for the variables. We design the one-sheet and two-sheet capacitive Jaumann absorbers at low frequency and the three-sheet at high frequency. The 20-dB attenuation bandwidth obtained for the one-sheet and two-sheet capacitive Jaumann absorbers are respectively, from 10 to 77 MHz and, from 4 to 61 MHz. For the three-sheet capacitive Jaumann absorber the 20-dB attenuation bandwidth obtained is, from 0.8 GHz to 280 GHz.