Search

In state-space systems, since the minimum statistical sensitivity structures obtained by the equivalent transformations are generally completely dense, they have defects and the numbers of parameters are very large. If this point can be improved, realizations with lower sensitivity and fewer parameters are obtained. This paper presents a synthesis of new minimum statistical sensitivity structures with fewer parameters as one such realization. First, to compare sensitivities of various realizations, assuming that only noninteger elements in the coefficient matrices contain small variations, the statistical sensitivity is defined. Since the sensitivity is invariant under the orthogonal transformations, using the symmetry of the coefficient matrices of the balanced realizations which are the members of the minimum statistical sensitivity structures, it is shown that a minimum statistical sensitivity structure can be realized with fewer parameters by defining an appropriate orthogonal transformation matrix. The number of parameters required in the proposed structure depends on the pole-zero configuration of the transfer function. Finally, various realizations are compared using numerical examples. As the parameters decrease, the insensitive elements increase. Therefore, the realization proposed in this paper has the lowest sensitivity. [ABSTRACT FROM AUTHOR]

This paper considers the determination of the frequency of a sinusoidal signal from its sampled values. A method is described by which the effect of the noise is reduced and the frequency is determined accurately. The method forms the sample set composed of four samples from the signal with unknown frequency, and the binary representation of the estimation is determined successively by varying the sampling frequency. However, the noise in the signal may produce an error. One means to eliminate the effect of the noise is to determine one bit of the binary number based on several sample sets. This paper proposes the following method to reduce the estimation error for the frequency due to the noise. The binary number is partitioned into the upper and lower bits. The upper bits are determined by the property of the rum, and the lower bits are determined by the property of the ramdom-walk. In either case, each bit is determined by several sample sets. By the evaluation based on the bit error rate, the accuracy is improved by combining the forementiomed methods, compared with the majority decision. A comprehensive evaluation was also made by simulation, indicating that the r.m.s. error can be reduced to 2 percent for the signal-to-noise ratio of 20 dB. [ABSTRACT FROM AUTHOR]

This paper presents the new upper bounds of channel width required in multi- layer channel routing under a wiring rule in which horizontal wires must be wired in the odd-numbered layers and vertical wires in the even-numbered layers. In (M ≧ 4)-layer channel routing, it is proved the upper bound is equal to [(dc +k)/[M/2]], where dc is the channel density, and k is 9 when M = 5, 7 when M is even, and 13 when M is odd and greater than 5. These upper bounds are obtained by presenting the multiple-layer channel routing algorithms, which wire the channels in a column-by-column manner. [ABSTRACT FROM AUTHOR]

In the use of the three-dimensional data obtained from a stereo system, laser range finder, and other equipment, merge processing is important. This paper proposes a new merge method for range images, based on the volume-based technique. In the proposed method, in contrast to the conventional method, the voxels are adaptively segmented according to the curvature of the surface shape to be reconstructed. This helps to represent the geometrical shape efficiently, effectively utilizing the computation resources. In the framework of the proposed merge process, additional features such as color and the laser reflectivity can be added to the three-dimensional geometrical information, which helps to maintain sharp edges in the texture and to perform rendering and the texture mapping effectively. By examining the consensus for both the geometrical shape and the color in the whole framework, the merge is made robust to noise. In this paper, the system is described first, and then the results of application to several kinds of actual data are presented. © 2007 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 90(11): 50–60, 2007; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20353 [ABSTRACT FROM AUTHOR]

In sensor networks, many of the sensor nodes are often positioned in clusters, and there is a correlation between the data obtained at the various sensor nodes by observation. This kind of observational data is collected at fusion centers. In this paper, we study the observation precision at fusion centers from the viewpoint of the Kullback–Leibler (KL) information. In particular, we consider transmission path errors arising when transmission is made from each sensor node via wireless communication paths, as a factor in the deterioration of the KL information, or in other words, observation precision, and study its effects and methods for reducing it. In this paper, we study the KL information obtained in fusion centers when there have been transmission path errors in cases where the correlation between observational data is all equal, and where the correlation is random. We show that the deterioration of the KL information can be reduced in fusion centers by using the correlation of observational data in the estimates of error data. In addition, we present examples of estimating the correlation between observational data in fusion centers that use correlation, and of estimating error data using correlation. © 2007 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 90(11): 1–10, 2007; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20250 [ABSTRACT FROM AUTHOR]

This paper proposes a real-valued orthogonal periodic sequence of a period N=2ν derived from a real-valued shift-orthogonal finite-length sequence of length M=2ν+1. This paper also explains the principle of a fast correlation algorithm that efficiently executes periodic correlation processing for this real-valued orthogonal periodic sequence. The sidelobe of an aperiodic autocorrelation function for a real-valued shift-orthogonal finite-length sequence (length of M) is 0 except for the right and left ends of the shift. If the subsequent sequence of first values repeatedly overlap the final values of this sequence, a real-valued orthogonal periodic sequence of a period N=M-1 can be obtained. A real-valued orthogonal periodic sequence of a period N=2ν generated from real-valued shift-orthogonal finite-length sequence of length M=2ν+1 is obtained by convoluting partial sequences and based on that controls the number of multiplications and the number of additions to increment on the order of Nlog2N without using fast Fourier transformation. © 2007 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 90(10): 18–28, 2007; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20294 [ABSTRACT FROM AUTHOR]

This paper discusses a method for increasing the eigenmode SNR by using a space-time linear precoder designed on a maximum likelihood (ML) basis (henceforth called the space-time linear ML precoder). First, the effect of the length of the precoded signal generated in the ML precoder on the eigenmode SNR is studied. It is shown that the eigenmode SNR can be increased by varying the precoded signal length depending on the condition of the channel. Next, a transceiver configuration for increasing the eigenmode SNR by the space-time linear ML precoder is presented. In the proposed method, error propagation may be generated due to the transceiver configuration and the BER characteristics may be degraded, even if the eigenmode SNR is increased. In this paper, in order to reduce the effect of the error propagation on the BER characteristics, the relationship between the average length of the error propagation and the precoded signal length in the proposed configuration is presented. It is demonstrated that the effect of error propagation can be minimized by limiting the region of search for the precoded signal length. Finally, the effectiveness of the proposed method is verified by numerical simulation. © 2007 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 90(10): 1– 8, 2007; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20304 [ABSTRACT FROM AUTHOR]

A serious problem of fingerprint verification devices which are easily available on the market is how to handle ‘‘dry fingers,’’ which make the reading of fingerprints difficult and often result in incorrect verification. Dry fingers are frequently observed in the elderly and in housewives who are frequently engaged in wet work. In these cases, the observed fingerprint tends to become an eroded fingerprint image, with disconnected ridges. This paper proposes a method of handling eroded fingerprint images in which such images that will result in incorrect verification are automatically identified and the ridges are recovered rapidly. First, a method for image quality evaluation that consists of evaluating the noise in the fingerprint image is discussed and the effectiveness of the method is demonstrated. Then, simple ridge recovery is performed by using edge emphasis for images which are automatically judged to be eroded fingerprints. In the experiment, a fingerprint database was acquired in winter, when dry fingers are common. The acceptance rate for the right person (recognition rate) is found to be 94.8% when the rejection rate for other persons is set as 99.99%. When the ridges are recovered in fingerprint images whose quality is judged to be low by the image evaluation measure proposed in this paper, the recognition rate is improved to 96.2%. Furthermore, when tolerance to dry fingers is considered and verification is not performed for low-quality image judged to be difficult to recover, the recognition rate is improved to 97.2%. In environments in which few users have dry fingers, the recognition rate will be improved to 98.8% when the rejection rate for other persons is set as 99.99%. © 2007 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 90(10): 40– 53, 2007; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20345 [ABSTRACT FROM AUTHOR]

In this paper, the author explains the recent evolution of algorithms for minimum partitioning problems in graphs. When the set of vertices of a graph having non-negative weights for edges is divided into k subsets, the set of edges for which both endpoints are contained in different subsets is called a k-way cut or k-cut. The problem of obtaining the k-way cut that minimizes the sum of the weights is an important research topic that appears in many practical applications such as VLSI design. In this paper, the author introduces recent results including cases in which sets of terminals or sets of terminal pairs that are to be separated are further specified in this problem and cases in which the objects to be partitioned are extended from graphs to hypergraphs or submodular set functions. © 2007 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 90(10): 63– 78, 2007; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20341 [ABSTRACT FROM AUTHOR]

This paper considers a vehicle-borne target detection system for short range in which an UWB (Ultra Wideband) impulse radar is used. The UWB impulse radar uses a pulse with widths of less than nanosecond order. Therefore, many multipath returns are observed on the receiving side. An IPCP (Inter-Period Correction Processing) receiver that resolves complexity of the receiver caused by multiple paths has been proposed. However, the integration time of the correlator in the IPCP receiver is limited by the period of the transmission signal (code length) and affects the range measurement accuracy. This paper considers vehicle mounting for applications. In order to assure safety, a receiver is constructed with several IPCP receivers in parallel, so that the range measurement accuracy is improved. In addition, a receiver which reduces the complexity of threshold setting in the conventional IPCP receiver when several targets are detected is proposed. The outputs of these receivers are derived analytically and improvement of the range measurement accuracy is confirmed. Further, based on the analytical results, the output of each receiver is characterized and the relationship of the threshold, detection probability, and false error probability is derived. Finally, the characteristics of use of the receiver as a vehicle-mounted radar are evaluated by simulation. © 2007 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 90(7): 1–14, 2007; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20303 [ABSTRACT FROM AUTHOR]

In the ubiquitous computing environment, the physical locations of persons and goods are extremely important information. In this paper, the design of a fully distributed localization system based on ultrasound, mainly for the indoor environment, is described. This system performs localization with as few positioning references as possible by an iterative technique. When such a localization method is used, deterioration of localization accuracy due to no-line-of-sight signals and to accumulated errors is a problem. In this paper, the resolution of these problems is discussed and system evaluation by implementation experiments is carried out. With 24 devices, localization with an accuracy of about 20 cm is demonstrated. © 2007 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 90(6): 17– 26, 2007; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20249 [ABSTRACT FROM AUTHOR]

In this paper the authors propose a program allocation method for the purpose of achieving real-time processing using a chip multiprocessor-type data-driven processor. The data-driven processor can perform parallel processing of executable processes without interference as long as the pipeline is not overloaded. The turnaround time for the individual processes being subjected to multiprocessing is independent of other processes being executed at the same time, and is maintained as is for the execution time when a single process is executed. Therefore, real time for a program can be estimated, and if the program can be allocated to a process such that overload does not occur, then real-time processing can be achieved without scheduling during execution. In this paper the authors first formalize a program allocation method in a data-driven processor. They then propose a combined method using genetic algorithms and a heuristic algorithm that uses lists as a program allocation method that is better at satisfying the time constraints and offers greater scalability for large-scale program and multiprocessor systems to be subjected to allocation. Finally, the authors use experiments in which a communications protocol process is allocated to verify the effectiveness of their program allocation method. © 2007 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 90(6): 48– 60, 2007; Published online in Wiley InterScience (www.interscience. wiley.com). DOI 10.1002/ecjc.20324 [ABSTRACT FROM AUTHOR]

Pattern extraction based on the Hough transform has the excellent property that it is robust to noise and also to the occlusions in the image pattern. On the other hand, the Hough transform has an important problem as regards accuracy due to its discrete representation, which has long been discussed. This paper proposes the LMedS Hough transform, a new algorithm which uses a robust statistic as the criterion for the likelihood of the straight line. The method is not constrained by the majority principle of the conventional Hough transform, and it provides a possibility of pattern detection which differs from the Hough transform considered up to the present. It is noted that the LMedS Hough transform retains the characteristics of the conventional Hough transform, such as its global property. This paper describes the basic principles and details of the proposed algorithm and analyzes its performance from the theoretical and experimental viewpoints. © 2007 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 90(5): 31– 39, 2007; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20320 [ABSTRACT FROM AUTHOR]

In this paper, we describe a method for creating all-in-focus photographic images where all of the subjects are in focus as much as possible from partially in-focus photographs with only its subject in focus when photographed by a camera with manual focus. Usually, the focusing range and the depth of field are wide, but become extremely narrow for nearby subjects. Consequently, it is difficult to photograph many subjects in focus in closeup photography. In this paper, we can combine the transform coefficients using wavelet transforms and easily create photographic images with natural focus but without noticeable seams. © 2006 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 90(3): 57–66, 2007; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20236 [ABSTRACT FROM AUTHOR]

This paper presents a new framework for observation of sporting events using HMD by applying the method of free viewpoint video synthesis for dynamic events to mixed reality. According to the viewpoint position of an observer, a virtual view image is generated by view interpolation among multiple sporting videos captured in a stadium, and then overlaid onto the real world via HMD. This makes the observer feel as if the match is played in front of his/her eyes. The proposed method performs virtual view synthesis and geometric registration between the real world and the virtual view images using projective geometry between cameras, which can be estimated from correspondence of natural feature points. It does not require calibration of multiple cameras imaging the sporting match and HMD camera capturing the real world. In this paper, we have constructed a mixed reality presentation system for a soccer match in order to demonstrate the utility of the proposed method. This system can overlay soccer scenes captured with multiple cameras in a soccer stadium onto a desktop field model with HMD. [ABSTRACT FROM AUTHOR]

In this paper, we propose frequency assignment as a method for multichannel speech separation. In this method, the magnitudes of the absolute values of each component in the frequency domain are compared, then based on this comparison result determines to which channel a frequency component originally belongs. The speech is separated by assigning the frequency to the determined channel. Therefore, it becomes possible to separate multichannel speech with a low computational load. However, it has the restriction that each speaker must speak close to his microphone. For situations in which this premise does not hold, we present a solution that uses beam-forming toward the sound sources. We also show the application of the frequency assignment to sound source localization. In this paper, we demonstrate the usefulness of the proposed method through simulation tests and tests in a realistic environment. © 2006 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 90(1): 59–70, 2007; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.10165 [ABSTRACT FROM AUTHOR]

This paper considers an identification method of the relay transmission characteristics including multipaths and coupling effects at the relay stations of a single-frequency network in an orthogonal frequency division multiplexing (OFDM) system. Estimating the transmission characteristics over the entire frequency range of OFDM transmission paths is extremely difficult because the transmitted signals in an OFDM system are limited to some bandwidth. Therefore, an objective of this paper is to propose a new approach that uses the guard interval feature of the OFDM signal to identify the relay transmission characteristics, multipath characteristics, coupling wave characteristics, and sensitivity characteristic of a closed-loop system including the coupling wave canceler, and evaluate the identification errors. The effectiveness of the proposed method is demonstrated through numerical examples. © 2006 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 89(12): 10– 19, 2006; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20230 [ABSTRACT FROM AUTHOR]

In this paper, we will discuss a highly dependable system configuration method and recovery method for our proposed Fat-Btree structure, which is a directory structure for shared-nothing parallel computers. The goals are to enhance the service of operation during failure, to minimize the probability of data loss, and to enhance availability by combining physiological logging, logical logging, disk mirroring, staggered allocation, etc. Various system configurations formed by the combination of these methods will be quantitatively evaluated in this paper. As a result, it will be shown that a combination of physiological logging and disk mirroring is most appropriate when hardware cost can be ignored, and that a combination of global logical logging and physiological logging is most appropriate when hardware cost is considered. © 2006 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 89(12): 42–58, 2006; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20288 [ABSTRACT FROM AUTHOR]

This paper proposes a new method of similar image retrieval based on the spectral distribution of a multispectral image. The greatest advantage of multispectral imaging is that the color and reflectivity of the object can be determined accurately. In similar image retrieval these physical variables must be discriminated and accurately evaluated. However, in the conventional retrieval method, the color composition ratio is emphasized rather than evaluation of the color (as a physical variable). In the histogram intersection method, for example, which is a form of similarity calculation based on the histogram, the similarity is calculated from only the shape of the histogram, that is, the color composition ratio. Thus, the similarity may be understated simply because the composition ratio is different, which makes this approach unsuitable for multispectral image retrieval, in which the color similarity itself is to be evaluated. This paper proposes a method of retrieval which does not depend on the color composition ratio (a physical variable), but is based on the similarity of the physical variable itself. The method is based on the distance between images, which accurately represents the difference in the physical variables by vector quantization. The effectiveness of the proposed method is demonstrated by an experiment on similar image retrieval for multispectral images. © 2006 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 89(11): 19–29, 2006; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20273 [ABSTRACT FROM AUTHOR]

This paper discusses parameter estimation for a continuous density HMM (hidden Markov model) method of visual speech recognition. Past studies of visual speech recognition can be broadly divided into two approaches, the image-based method and the model-based method. The image-based method is a method in which some preprocessing, such as subsampling and principal component analysis, is applied to the pixel values of the original image, and the result is used as the feature vector. In this approach, the position and size of the lips and the illumination conditions have a direct effect on the recognition rate. Thus, the normalization of these factors is the basic technique. The ordinary approach in the conventional normalization is to provide a criterion independently of the HMM, and to apply normalization before learning. In this paper, normalization by the ML (maximum likelihood) criterion is considered. Normalized training is proposed in which the normalization processes for elements such as the position, size, inclination, mean brightness, and contrast of the lips are integrated with the training of the model. The proposed method is formulated on the basis of an EM (expectation maximization) algorithm in which monotonically increasing behavior of the likelihood of the training data is guaranteed, by iteration of normalized training. The effectiveness of the proposed method is demonstrated in a word recognition experiment using the M2VTS database. © 2006 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 89(11): 40–50, 2006; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20281 [ABSTRACT FROM AUTHOR]

This paper proposes a microphone array unified with an echo canceller that resolves the problem of deteriorated echo suppression performance due to variations of the echo path when a microphone array and an echo canceller are connected. In a teleconferencing system, high-quality sound acquisition with noise suppression is desirable in addition to acoustic echo suppression. However, in a configuration in which a microphone array to suppress the noise and an echo canceller to suppress the echo are simply connected, the adaptation for the echo canceller cannot follow the echo path variations due to directivity control of the microphone array, so that there is a problem of deteriorated echo suppression performance. In the microphone array with an echo canceller proposed in this paper, the training of the microphone array and that of the echo canceller are performed in a unified way. In this way, the problem of deteriorated echo suppression performance due to echo path variations can be resolved. An evaluation experiment shows that the proposed method can perform stable noise suppression and echo cancellation. © 2006 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 89(10): 23–32, 2006; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20090 [ABSTRACT FROM AUTHOR]

This paper is concerned with the lowering of spatial resolution by γ ray attenuation and the collimator aperture, which are serious factors degrading the quantitative performance of SPECT (single-photon emission CT). The limit of analytical correction is investigated from the viewpoint of the aperture angle and noise. The methods considered in this paper are the analytical attenuation correction proposed by Kudo and Saito and the analytical aperture correction based on the stationary phase principle proposed by Lewitt and colleagues. Both of these methods are intended to correct projection data affected by attenuation and aperture blur in the Fourier space. Simulation results show that the correction for the attenuation and the aperture is effective if the single-sided aperture is up to 1°. It is also shown that the statistical noise should be carefully handled in applying the correction. © 2006 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 89(7): 42–50, 2006; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20253 [ABSTRACT FROM AUTHOR]

In this paper we propose a method for the reciprocal use of knowledge between association rule extraction (a technique from data mining) and HMM-based information extraction. Association rule extraction is a method that is generally applied to structured data and is either inefficient or intractable when applied to data that lacks an accurate information schema. In this paper we use HMM-based information extraction to automatically recognize the attributes of each word in a set of textual data and assign tags in order to structure this data. In addition, we use the degree of association between words seen in the association rules extracted from the structured data to weight the HMM parameters in order to improve the recognition accuracy of the information extraction procedure and allow knowledge to be reciprocally shared between the data mining and information extraction procedures. Besides this we propose a data mining method that uses representative schema patterns found in different types of semistructured data such as tagged or BibTeX data using schema discovery methods. © 2006 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 89(7): 51–60, 2006; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20256 [ABSTRACT FROM AUTHOR]

The bank-based multiport memory is a better composition approach to realizing realistic chip area and high access bandwidth than a conventional N-port memory cell approach. However, this method is unsuitable for large numbers of ports and banks because the hardware resources of the crossbar network which connects the ports and banks increase in proportion to the product of the numbers of ports and banks. In order to solve this problem, this paper proposes a new bank-based multiport memory architecture using a blocking network instead of a crossbar network. Many blocking networks have been researched so far. However, these researches evaluated hardware resources based on the number of switches, but the compositions and circuit scale of the switches used in crossbar network and blocking network are different. Hence, this paper compares the number of transistors to show that the bank-based multiport memory using the blocking network achieves high access bandwidth with smaller hardware resources than the conventional approach. According to our results, our approach achieves the same access bandwidth with half the number of transistors, for 512 ports and 512 banks. © 2006 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 89(6): 22–33, 2006; Published online in Wiley InterScience (www. interscience.wiley.com). DOI 10.1002/ecjc.20205 [ABSTRACT FROM AUTHOR]

It often happens in medical image processing that shadows with very weak contrast must be handled. When an ordinary difference filter is applied, shadows with a strong contrast other than the target shadow give a disturbance signal, making it difficult to extract the desired region embedded in the disturbance-producing shadows. This paper notes the directions of the gradient vector in the image and discusses a filter which can detect objects with high sensitivity based on the direction distribution. In this paper, the convergence index of the vector is generally defined, and a general expression for the convergence index filter which can be applied to both 2D and 3D images is presented. Specifically, the iris filter and the line convergence index filter are considered, and their responses are analyzed, together with their relations. An application to the actual medical image is presented, demonstrating the effectiveness of the idea. © 2006 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 89(6): 34–46, 2006; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20247 [ABSTRACT FROM AUTHOR]

This paper describes a method of configuration of a complex RiCR filter with coupled Chebyshev characteristics. The frequency shift method is a well-known method of obtaining a complex filter from a real filter, in which the transfer characteristic of the real filter is shifted along the frequency axis. The complex coupled Chebyshev filter obtained by the frequency shifting of a coupled Chebyshev filter has an excellent attenuation characteristic, but the number of constituent elements for active realization is large and the circuit is complex. In the proposed method, the complex coupled Chebyshev filter is first transformed to a complex RiCR filter so that the number of constituent elements for active realization is reduced and the circuit is simplified. In this paper, a method is presented for transforming the complex coupled Chebyshev filter to a complex RiCR filter. The attenuation characteristics and the numbers of constituent elements of the complex filter designed by the proposed method, the complex coupled Chebyshev filter, and the conventional complex RiCR filter, are compared for confirmation of the effectiveness of the proposed method. © 2006 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 89(5): 12–20, 2006; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20212 [ABSTRACT FROM AUTHOR]

This paper considers an input method for the wearable computer environment using a wearable video camera. An input system is proposed in which the user performs the motion of writing an alphanumeric character in the air, for input to the computer. In the proposed method, the user's hand motion is imaged by a wearable video camera, and letter input is performed by analysis of the monochrome gray-level picture in the computer. When a letter is written in the air, it is difficult to identify the start and end of the letter input by the user, or the start and end points of the segments composing the letter. Furthermore, in a wearable computer environment it is desirable that the hand motion be detected in the background, which is continually changing both in the daytime and at night. Consequently, this paper proposes the following three procedures: (1) extraction of the user's hand region in the picture frame by visible ray or infrared illumination; (2) determination of the center of gravity of the user's hand motion in the air, using the intensity difference between picture frames as a clue; (3) continuous DP matching to identify the letter. In order to solve the essential problem of letter recognition in the air, a writing format for alphanumeric characters is proposed. The proposed system is implemented on a video camera and a laptop PC, and an experiment on writing of letters in the air is performed with five subjects and 360 alphanumeric characters. A recognition rate of approximately 75% is obtained. © 2006 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 89(5): 53–64, 2006; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20239 [ABSTRACT FROM AUTHOR]

This paper discusses codes, and a decoding method, that can correct compound errors containing a mix of random and burst errors. Specifically, the paper provides a minimum compound distance that represents a compound error correcting capacity for iterated codes, proposes a compound error correcting decoding algorithm for the code, and demonstrates that the decoding algorithm is capable of correcting compound errors up to half the minimum compound distance of the code. The proposed decoding algorithm is a generalized decoding method in which the decoding method for iterated codes indicated by Reddy and Robinson is a particular case. By this generalization, the proposed decoding algorithm can correct compound errors that cannot be corrected by conventional decoding methods. © 2005 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 89(1): 60–72, 2006; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20035 [ABSTRACT FROM AUTHOR]

This paper considers an Internet agent system in which a tremendous number of agents operate, frequently appearing and disappearing, and discusses the fault-tolerant algorithm. Application of the snapshot algorithm to the agent system is considered. The snapshot algorithm is used to view the whole situation (snapshot) of the distributed system. The snapshot algorithm of Chandy and Lamport [2] is considered as a representative snapshot algorithm, in terms of the high efficiency and the simplicity of the procedure. It is not practical, however, to apply their snapshot algorithm to the distributed agent system in which a tremendous number of agents operate. From such a viewpoint, this paper extends the idea of Chandy and Lamport's algorithm and proposes a subsnapshot algorithm, in which the snapshot is taken among the agents who are in the causal relation, through message exchange and agent creation. Then, an efficient rollback algorithm is proposed, which is based on the snapshots taken by the subsnapshot algorithm. In the general rollback algorithm utilizing the snapshot, all agents must roll back. In contrast, in the rollback algorithm proposed in this paper, it suffices that only some agents should roll back. © 2005 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 88(12): 43–57, 2005; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20208 [ABSTRACT FROM AUTHOR]

As a measure for preventing eavesdropping on land-mobile communication, this paper studies an application of a private key agreement based on the propagation characteristics of channel to a practical system without key delivery. The paper proposes a private key agreement based on the time-varying frequency characteristics suitable for OFDM. In the proposed method, the time-varying frequency characteristics are alternately measured on the basis of the channel reciprocity and generation of the key and key agreement is attempted. Also, in the proposed method, for prevention of errors in key generation, measurement time difference compensation of the channel, noise reduction by the synchronous addition process, and key disagreement correction by an algebraic decoding method are applied. Further, in order to evaluate the safety of the key, a method for evaluation of independence of the key is discussed. As a system mode for evaluation of the performance of the proposed method, a model is constructed for study that is based on the specification of the IEEE802.11a wireless LAN using the OFDM, so that verification of an operation can be carried out in a more realistic environment. Numerical simulation indicates that it is possible to share the private key in a realistic setting with an SN ratio of 15 dB by means of various measures for prevention of key disagreement. Also, an evaluation confirms that independence is sufficiently maintained. © 2005 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 88(9): 1–10, 2005; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20167 [ABSTRACT FROM AUTHOR]

It is conceivable to embed additional information with a view to judge whether certificates existing as paper media or digital image data have been altered. The authors have already proposed and discussed a method for configuration of image information enabling reconstruction of the original image information from contaminated image information on paper by means of coding theory. Based on this proposal, a new method is being proposed for effective detection of alteration of digital image data by applying coding theory and the cryptographic theory at the same time. In the present paper, this method is critically discussed and a new method is proposed. © 2005 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 88(8): 9–17, 2005; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20150 [ABSTRACT FROM AUTHOR]

In most of the methods of public key cryptography devised in recent years, a finite field of a large order is used as the field of definition. In contrast, there are many studies in which a higher-degree extension field of characteristic 2 is fast implemented for easier hardware realization. There are also many reports of the generation of the required higher-degree irreducible polynomial, and of the construction of a basis suited to fast implementation, such as an optimal normal basis (ONB). For generating higher-degree irreducible polynomials, there is a method in which a 2m-th degree self-reciprocal irreducible polynomial is generated from an m-th degree irreducible polynomial by a simple polynomial transformation (called the self-reciprocal transformation). This paper considers this transformation and shows that when the set of zeros of the m-th degree irreducible polynomial forms a normal basis, the set of zeros of the generated 2m-th order self-reciprocal irreducible polynomial also forms a normal base. Then it is clearly shown that there is a one-to-one correspondence between the transformed irreducible polynomial and the generated self-reciprocal irreducible polynomial. Consequently, the inverse transformation of the self-reciprocal transformation (self-reciprocal inverse transformation) can be applied to a self-reciprocal irreducible polynomial. It is shown that an m-th degree irreducible polynomial can always be generated from a 2m-th degree self-reciprocal irreducible polynomial by the self-reciprocal inverse transformation. We can use this fact for generating 1/2-degree irreducible polynomials. As an application of 1/2-degree irreducible polynomial generation, this paper proposes a method which generates a prime degree irreducible polynomial with a Type II ONB as its zeros. © 2005 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 88(7): 23–32, 2005; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20151 [ABSTRACT FROM AUTHOR]

This paper proposes a new algorithm for the unconstrained optimization problem, which is derived as an application of the control theory called Direct Gradient Descent Control. The static optimization problem is solved by using a dynamic controller in which the convergence speed can be greatly accelerated. The major idea is to consider the objective function F(x) and its time derivative dF(x(t))/dt as the evaluation criteria, and to apply the gradient descent method. It is verified by simulation that the proposed dynamic mathematical programming has a very great capability to converge to the optimal solution. It is also interesting to note that the method has to a certain extent the ability to find not only local optimal solutions, but also global optimal solutions. This paper improves the basic algorithm of dynamic mathematical programming so that it is suited to global optimization. The new algorithm generates movement toward the global optimal solution by utilizing two different trajectories with interaction (the chaotic trajectory and convergence trajectory). It is verified by simulation that the new algorithm functions as a global optimization procedure with a high success probability. © 2005 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 88(6): 20–27, 2005; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20094 [ABSTRACT FROM AUTHOR]

This paper proposes a new lattice structure for the M-channel paraunitary filterbank. When a filterbank is realized by using a lattice structure, there is an advantage that properties such as phase linearity and the paraunitarity can be retained if the coefficients are quantized. A large number of lattice structures have been proposed, but there remains the problem that the proposed structures contain redundancy. This paper first eliminates the redundancy from the lattice structure and presents a new lattice structure for the generalized filterbank without restrictions such as phase linearity. Based on the proposed structure and restricting the parameters, a pairwise mirror image filterbank structure is presented. Finally, the usefulness of the approach is demonstrated by simulation. © 2005 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 88(6): 12–19, 2005; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20125 [ABSTRACT FROM AUTHOR]

In this paper, the various protocols used to achieve node mobility on the Internet are compared and evaluated. Node mobility can be defined as hiding from the application or user changes in the connection point to the Internet of the terminal. Node mobility can be broken down into two forms of mobility: wide area mobility and local mobility. Wide area mobility can be achieved by satisfying two factors: mobile node reachability and communication continuity. Local mobility can be achieved by satisfying two factors: packet loss avoidance and control traffic suppression. In this paper, the various proposed protocols for achieving node mobility are analyzed, and the methods used to satisfy the above factors are classified. Each classified method is then analyzed and evaluated from the standpoint of communication efficiency, scalability, security, fault tolerance, and ease of implementation. In addition, specific examples are given for the representative protocols. © 2005 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 88(6): 39–59, 2005; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20173 [ABSTRACT FROM AUTHOR]

An acoustic echo canceler, needed for hands-free telecommunications in videoconferencing, has been realized with sufficient performance for use in a quiet meeting room. However, demand has been increasing for the use of hands-free teleconferencing in an environment with ambient noise such as conference calls in a noisy office or in a running automobile. In response to such demands, this paper proposes an echo canceler with a capability of simultaneously suppressing stationary ambient noise superposed on the speech and also residual echo, which cannot be eliminated by conventional echo cancelers because it is buried in noise. In this paper, the noise reduction and echo reduction process, based on the short-time spectrum amplitude estimation, is first explained. Then the results of objective evaluation of a test-fabricated echo canceler are presented. The evaluation results confirm that the echo can be reduced by about 30 dB and the ambient noise by about 15 dB even in an environment with ambient noise, allowing satisfactory hands-free conference calling to be realized. © 2004 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 88(3): 21–31, 2005; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20123 [ABSTRACT FROM AUTHOR]

This paper studies an adaptive filter with a restricted linear sum of the moments of tap coefficients and investigates its applications. In general, when a sampled value filter or a digital filter represented by a surface acoustic filter is realized, the maximum absolute values or the variances of the tap coefficients are restricted. In this paper, restrictions on the maximum values or the variances are imposed by means of a restricted linear sum of several moments around an arbitrary positive number, facilitating the realization of the filter. The number of the moments and the weights are determined in such a way that the frequency characteristics of the filter after restriction satisfy the given specifications. The adaptive system consists of an ideal filter without restriction on the tap coefficients and an adaptive filter placed in parallel. The adaptive filter consists of a transversal filter. Its adaptive algorithm is a gradient algorithm with a small amount of processing, with a restricted linear sum of the moments of the tap coefficients. Since the proof of uniqueness of the converged value of the adaptive algorithm and the upper limit of the loop gain in the algorithm are formulated, the filter designer can easily obtain filter parameters with restricted tap coefficients. Finally, verification of the discussed results is rigorously carried out by computer simulations. © 2004 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 88(3): 11–20, 2005; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20126 [ABSTRACT FROM AUTHOR]

The two-dimensional wavelet transform usually used applies one-dimensional wavelet transforms in the vertical and horizontal directions. By carrying out the wavelet transform, the local frequency can be analyzed and manipulated. However, when two-dimensional extension is performed by means of one-dimensional transforms in the vertical and horizontal directions, the components at 45° and 135° cannot be distinguished. The method proposed in this paper allows local processing corresponding to four or six directions by combining the wavelet transform and perfectly reconstructable filters. In this paper, the proposed method is described and its effectiveness is demonstrated by simulation of noise elimination. © 2004 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 88(3): 1–10, 2005; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20129 [ABSTRACT FROM AUTHOR]

This paper discusses a routing scheme for an autonomous distributed wireless network (mobile ad-hoc networking) using an ultra-wideband (UWB) system. The UWB system is a kind of spread-spectrum communication system which has the function of measuring the distance between the transmitter and the receiver with high accuracy. Using this distance measurement function, the received energy per bit is kept constant by adjusting the spreading factor according to the transmission distance. This paper considers a packet-forwarding scheme based on the above feature of the UWB system, where the spreading factor is adjusted according to the distance between the transmitter and the receiver. The objective is to realize higher throughput than in single-hop transmission by using multihop transmission which controls the maximum transmission distance. © 2004 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 88(2): 22–30, 2005; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20082 [ABSTRACT FROM AUTHOR]

In this paper, in order to improve the packet error rate of a ubiquitous antenna-based SDMA (Space Division Multiple Access) system, the application of a successive interference canceler (SIC) is proposed. In a ubiquitous antenna-based SDMA system, radio base stations (RBS) distributed throughout the service area are connected by radio-on-fiber links to the central control station (CCS). By performing all signal processing functions such as wireless modulation and demodulation at the CCS, simultaneous reception of several signals transmitted in the same frequency range is made possible and hence the frequency utilization efficiency can be enhanced. When an SIC is applied to the ubiquitous antenna system, the magnitude relationship of the power received by the user is different for each RBS as the RBSs are distributed at several locations. Hence, strength comparison of the received power is not easy. In the SIC proposed in this paper, the SINR (Signal to Interference plus Noise Ratio) after MMSE is estimated, and then the interference is eliminated according to size. A computer simulation is performed assuming a QPSK-OFDM (Quadrature Phase Shift Keyed–Orthogonal Frequency Division Multiplex) signal with a coding rate of 1/2 in accordance with the IEEE 802.11a specification. It is made clear that by the use of the proposed scheme the required transmitted power satisfying the requirement of an average BER of 10-4 can be reduced by about 7.2 dB, and that with a packet transmission success rate of 0.9 it can be reduced by 8 dB. © 2004 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 87(12): 10–25, 2004; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20076 [ABSTRACT FROM AUTHOR]

This paper considers a technique for prerecorded TV programs in which captions for the hearing-impaired are automatically superimposed on the basis of the program VTR and the advance electronic script. A method of detecting the caption presentation timing by detecting the synchronization timing for the speech and captions is described. For broadcast speech on which background sound is superimposed, it is difficult to achieve high detection accuracy by timing detection based only on a phoneme HMM word spotter. Consequently, this paper proposes the following method. For each sentence in the program script, multiple timing candidates are detected by word spotting. The optimal timing for the whole program is selected by using dynamic programming based on three scores (the time-order in the manuscript, the ratio of the utterance times estimated from the manuscript, and the likelihood of being speech) in addition to the acoustic likelihood. An evaluation experiment was performed on 10 sessions of a documentary program. Assuming tolerable detection errors of 1 and 3 seconds, detection rates of 99.0 and 99.7% were obtained, respectively, indicating that the method is of practical value. © 2004 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 87(11): 73–82, 2004; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20121 [ABSTRACT FROM AUTHOR]

This paper considers the recognition of moving images and proposes a new framework in which the local features, the global structure, and motion information are handled comprehensively. The method is applied to the extraction of facial expressions and its effectiveness is demonstrated. In most conventional moving image recognition methods, recognition is performed on the basis of the sequence of recognition results for individual frames. However, the authors believe that the motion information should be more positively utilized in the recognition of the individual frames. When the time-series data are combined into the recognition of a still image, however, the amount of information is tremendous, making the processing complicated. The method proposed in this paper extends the concept of the labeled graph matching method, which has been used for still images, to moving images. The proposed method handles sparse graphs and can prevent an increase in the amount of computation. By dynamically adjusting the features to be handled according to the stage of processing, complex processing in dynamic image recognition can be integrated in a simple and straightforward way. As practical examples, the human head, and parts of the head, are extracted, indicating the effectiveness of the proposed method. © 2004 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 87(10): 35–43, 2004; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20106 [ABSTRACT FROM AUTHOR]

This paper discusses selective stimulation, which is a problem common to various electrical stimulation techniques. The induction of nerve activity by electrical stimulation is a technique which is widely used in medical and health care. In order to reduce unpleasant feelings during stimulation and to diversify the induced sensations, it is necessary to control the stimulus waveform and to design the electrode in appropriate ways. From the viewpoint of nerve stimulation, the above requirement can be interpreted as a selective stimulation problem in which the target nerve fiber is stimulated and the activities of other nerves are minimized. At the present stage of electrical stimulation, however, there is no means of representing selectivity of stimulation mathematically, although precise modeling of the stimulus site is possible. Consequently, the design for selective stimulation is based on trial and error. The purpose of this paper is to construct a mathematical framework for the design of selective stimulation. The requirement for selective stimulation is formulated as a quadratic programming problem which gives a route to the optimal stimulation technique. Several simple examples are solved, and the usefulness of the proposed method is shown through comparison to selective stimulation techniques that were proposed in the past. © 2004 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 87(9): 62–72, 2004; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20098 [ABSTRACT FROM AUTHOR]

This paper shows a technique for efficiently analyzing the information flow of a program that generally includes recursive procedures. This technique enables security levels representing the secrecy of data to be assigned according to an arbitrary finite distributive lattice. This paper proves the soundness of the proposed information flow analysis method and shows that this algorithm can be executed in cubic order of the program description size. In addition, a congruence relation is used to extend the algorithm so that functions can be appropriately introduced to enable information within arguments to be hidden. Finally, a prototype system is used to present execution results of the proposed algorithm. © 2004 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 87(9): 48–61, 2004; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20099 [ABSTRACT FROM AUTHOR]

This paper proposes a method of realizing a handwritten mathematical formula input system that works in real time. In the handwritten character input, the data is input generally for each stroke (segment). The proposed method is based on this property. The mathematical formula is recognized on the basis of the result of recognition of the individual input strokes and their positional relations. Then the LaTeX source code is generated. In this paper, mathematical symbols are considered as characters, and are integrated into character recognition. More precisely, each time a stroke is input, it is recognized as a single-stroke character. The crossing with the already input characters is examined, and the character is extracted so that the likelihood of forming character is maximized. Based on the positional relations among the extracted characters, the structure of the mathematical formula, such as superscripts/subscripts and fractions, is recognized. The mathematical structures that can be recognized by the proposed method are the structures obtained by combining five different structures, namely, superscript/subscript, root, overline, underline, and fraction. The recognition is not affected by the order of input. The proposed method is applied to various mathematical formulas, and it is seen that the proposed system works adequately as an input environment for real-time processing. © 2004 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 87(8): 66–81, 2004; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20086 [ABSTRACT FROM AUTHOR]

This paper proposes a multiple adaptive notch filter with the p-th power average of the estimated error E[ep(n)] |as the cost function. The adaptive filter consists of a cascade connection of second-order adaptive notch filters composed of the allpass filters. In general, the elimination bandwidth of the notch filter is desirably small from the practical point of view. However, as the elimination bandwidth is set smaller, the convergence speed of the tap coefficients of the adaptive filter becomes much slower. From the point of view of convergence theory, the radius of the pole of the second-order allpass filter needs to be close to 1, so that the elimination bandwidth must be extremely small. Convergence speed must be sacrificed due to the practical and theoretical points of view. In this paper, an adaptive notch filter is studied in which the elimination bandwidth of the notch filter is designed small while the gradient of the tap vector of the cost function used in the adaptive algorithm is made effectively equal to the case in which the elimination bandwidth is rather large. In this way, realization of a multiple adaptive notch filter with a high convergence is possible. The absolute value of the transfer function of the notch filter is close to 0 near the elimination frequency while it is close to 1 at other frequencies. Hence, by setting the power of the absolute value larger, the elimination bandwidth becomes larger and the smoothness near the elimination frequency is increased. By means of the gradient algorithm with a cost function E[ep(n)], this property can be utilized so that excellent convergence speed and steady-state characteristics can be obtained. Finally, verification of the analytical results and comparison with the conventional method are carried out by computer simulation. © 2004 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 87(8): 1–8, 2004; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.10162 [ABSTRACT FROM AUTHOR]

In a workflow information processing system, usually optimization of the workflow overall and design of each process which comprises the workflow are performed separately. As a result, inefficient procedures and duplication of processes remain in the workflow. In order to resolve this problem, the authors apply activity-based costing management (ABM), a managerial accounting method, to workflow design in this paper. Note, however, that under a conventional ABM which allocates cost and time, sufficient analytical results cannot be obtained. Thus, in this paper the authors propose a new analysis method which allocates data items in the system. Under this method, a data item in the system or an operator process is defined as “information,” and then classified as value-added information or non-value-added information depending on whether or not it contributes to the manufacturing resource produced by the system. The authors apply their proposed method to the workflow of the research paper manuscript system at the Institute of Electronics, Information, and Communication Engineers and confirm that it reduces the number of steps in the workflow by 12%. © 2004 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 87(8): 55–65, 2004; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.20087 [ABSTRACT FROM AUTHOR]

This paper proposes a relaxation algorithm for an inverse filter that is adaptive to environmental change in a multichannel sound reproduction system. For a conventional multichannel sound reproduction system, a design method has been proposed for an inverse filter by means of the Moore–Penrose-type generalized inverse matrix. However, in the inverse filter in the conventional method, background noise and line noise may be enhanced when the room transfer characteristics vary. In order to resolve this problem, an adaptive relaxation method based on the truncated singular value decomposition (TSVD) has been introduced. In the present paper, a microphone for monitoring the noise is placed at a location far from both ears of the listener. Relaxation of the inverse filter is carried out by using the reproduced sound observed by this microphone. From the results of computer simulation, it is found that the proposed method prevents amplification of the noise. Also, as a result of a subjective evaluation experiment on sound reproduction in a real acoustic environment, the sound quality was found to be improved significantly without damaging the sound localization of position. © 2004 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 87(7): 15–26, 2004; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.10159 [ABSTRACT FROM AUTHOR]

A real-valued shift-orthogonal finite-length sequence is a sequence for which the side-lobe of the aperiodic autocorrelation function is zero except at the two ends. It is effectively used in spread-spectrum communications and other applications. This type of sequence can be synthesized by using element sequences of adequate lengths. Consequently, fast correlation processing can be realized by dividing the process for each element sequence. A problem in this sequence is that the round-off error of its real values has an effect on the correlation characteristics. Therefore, this paper investigates the effect of the round-off error on integrated correlation processing and divided correlation processing. In the former, the mean-square error of the correlation output is inversely proportional to the sequence length. In the latter, it is the sum of the mean-square errors of the element sequences, which are inversely proportional to the respective sequence lengths. Thus, we see that the error of shorter element sequences should be reduced in the latter method in order to speed up correlation processing. This paper considers in particular a method by which the mean-square error is reduced without significantly enlarging the circuit scale. © 2004 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 87(6): 1–12, 2004; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.10152 [ABSTRACT FROM AUTHOR]

This paper considers the continuous distribution HMM and discusses speed improvement by parallel processing and VLSI design, aiming at the realization of a high-speed speech recognition LSI. The continuous distribution HMM has high recognition performance. But a large amount of computation is required for the calculation of the output probability, and speed improvement is highly desirable. Regarding speed improvement of the output probability calculation, this paper proposes blockwise parallel processing and an architecture for performance. The minimum optimal bit width is investigated for recognition by fixed-point operations. Based on the proposed architecture, VLSI design for an HMM is performed and the result is evaluated. It is shown that the proposed architecture has sufficient power to construct a recognition system that can respond in real time, requiring 3 ms for the recognition processing of 100 words. © 2004 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 87(5): 12–23, 2004; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.10099 [ABSTRACT FROM AUTHOR]