電気学会論文誌C(電子・情報・システム部門誌)
Online ISSN : 1348-8155
Print ISSN : 0385-4221
ISSN-L : 0385-4221
124 巻, 7 号
選択された号の論文の26件中1~26を表示しています
特集:平成15年電気学会電子・情報・システム部門大会
特集論文
<光・量子エレクトロニクス>
  • 川瀬 晃道, 渡部 裕輝, 小川 雄一, 伊藤 弘昌
    2004 年 124 巻 7 号 p. 1339-1344
    発行日: 2004年
    公開日: 2004/10/01
    ジャーナル フリー
    We have developed a novel basic technology for terahertz (THz) imaging, which allows detection and identification of chemicals by introducing the component spatial pattern analysis. The spatial distributions of the targets are obtained from terahertz multispectral transillumination images, using absorption spectra measured with a widely tunable THz-wave source. First, the two kinds of chemical pellets were clearly distinguished. Next, we separated the component spatial patterns of frequency-dependent absorptions in chemicals and frequency-independent components such as plastic, paper and measurement noise in THz spectroscopic images. Hereby we expect this technology to contribute to the following potential applications: screening letters and packages, frisking, quality inspection of pharmaceuticals, and pathologic diagnosis.
  • 小川 博司, 鎌田 雅夫, 西尾 光弘, 増田 正孝, 郭 其新, 田中 徹, 近藤 祐治, 林田 和樹, 本岡 輝昭
    2004 年 124 巻 7 号 p. 1345-1351
    発行日: 2004年
    公開日: 2004/10/01
    ジャーナル フリー
    Saga Synchrotron Light Project is proceeding on schedule by the Saga prefecture government with the academic advises from Saga University. The present status of this facility is introduced briefly. At present, the building is finished and the 1.4 GeV ring is under construction at Tosu city in Saga prefecture. In parallel with construction of light source, three beam lines are planning by the Saga prefecture government with contribution of several universities in Kyushu area, e.g., Saga University. The outline of one of the three beam lines, which is for development of advanced materials and processing, is also reported.
<医用電子・生体工学>
  • 永岡 隆, 内山 明彦
    2004 年 124 巻 7 号 p. 1352-1358
    発行日: 2004年
    公開日: 2004/10/01
    ジャーナル フリー
    We developed wireless 3-D position sensor. Although we reported a wired position sensor, in order that the wire might affect movement of the sensor, we develop a wireless position sensor newly. For inspection of position measurement algorithm, we first developed a trial product of wireless capsule type sensor without regard to size. The position of the sensor was detected by using mutual induction. In order to maintain the electromotive force induced at secondary coils in the sensor constant, the current flowing into the primary coil was regulated using feedback control. The position of the sensor was measured stably using the feedback control, even when the sensor was relatively far from the primary coil. Induced electromotive forces are modulated to FM signals by using astable multivibrator and the FM signal is flowed into a living body directly with low current, and demodulated from electrode on body surface. The position of the sensor was determined from the current flowing into the primary coil. As a result of accuracy evaluation of the sensor in vitro, the error was 2.17±0.99 mm (Ave±S.D.) in a range of up to 1 m from the magnetic field generator. This sensor is applicable to medical capsule devices, such as capsule endoscope.
<通信・ネットワーク>
  • 園田 潤
    2004 年 124 巻 7 号 p. 1359-1366
    発行日: 2004年
    公開日: 2004/10/01
    ジャーナル フリー
    This paper describes the study of a fast electromagnetic (EM) wave propagation analysis that can solve electrically large domains using finite difference time domain (FDTD) method on cluster of personal computers (PC cluster). It reports an implementation of parallel FDTD using an MPI library on PC clusters of the computer system for education. Use of this method demonstrates that the speed-up ratio achieved for problem size 1200 × 1200 is about 55.0 using FDTD on 80 PCs. And also, indoor propagation of UWB pulse on the floor (1095.4λ× 98.6λ) is analyzed by the parallel FDTD using 40 PCs, computational time and memory have been reduced by 1/36.4 and 1/39.9, respectively. The results demonstrate that the parallel FDTD using PC cluster can analyze electrically large problems low computational costs than novel FDTD.
  • 白川 晃, 植田 憲一
    2004 年 124 巻 7 号 p. 1367-1374
    発行日: 2004年
    公開日: 2004/10/01
    ジャーナル フリー
    The recent progress of high-power fiber lasers is remarkable. Advantages of a fiber laser in characteristics such as ultra-low loss, high-damage threshold, and waveguide structure enable the outstanding high-power operation with excellent beam quality. In this paper, with the overview of the progressive advances of fiber lasers, our work on high-power fiber lasers is reported. The power limitation of double-clad fiber laser is discussed, which clarifies the importance of scaling the launching capability of pump power. A concept of a fiber-disk laser with a side-pumping geometry was developed and more than 1 kW output power was demonstrated, which is the highest power from a single fiber laser. For further power and brightness scaling, coherent beam combining of many fiber lasers is also investigated. N fiber lasers simply coupled by fused fiber couplers generate power-summed output from one of N fiber ports with high addition efficiencies of more than 85% for N≤8 due to selective excitation of constructive interference modes.
  • 川上 康人, 都築 伸二, 和崎 賢, 山田 芳郎
    2004 年 124 巻 7 号 p. 1375-1381
    発行日: 2004年
    公開日: 2004/10/01
    ジャーナル フリー
    Although power-line communications are expected as a promising method to realize an in-home network, one of their difficulties is an unnecessary radiation due to the electrical unbalance of power-line channels. To solve the problem, the statistical data of electrical unbalance based on the measurement at the practical fields, i.e., many and various Japanese residential power-lines, must be collected as the first step. In this paper, a method to produce a longitudinal-conversion-loss˜(LCL) probe that can be used for the practical measurement is studied. The LCL is a measure of the electrical unbalance, and is often used to determine the impairment of EMC performance of a system by the cabling. To improve the dynamic range of the probe, an adjustment method of transformers’ parameters is proposed. According to our scenario, the obtained dynamic range was better than that of the conventional probe with the frequency band higher than 2MHz. The achieved improvement was 6dB at 6MHz, and 14dB at 30MHz, for example.
<音声画像処理・認識>
  • Yang Jianhua, Yasutaka Fujimoto
    2004 年 124 巻 7 号 p. 1382-1389
    発行日: 2004年
    公開日: 2004/10/01
    ジャーナル フリー
    A normal expansion method combined with MDL (Minimum Description Length) principle is proposed to construct an approximation of general distribution using its discrete data. Each of data is transformed to be a basic normal distribution and their composition is used to approximate the original distribution. Theoretically the composition is the same as the original distribution if the number of discrete data is infinite. For the finite quantity of data, a suitable approximation can be done if a variance is selected for each basic normal distribution to reach the MDL value. In this paper, the original distribution with single peak such as normal distribution, beta distribution etc., is firstly investigated and generated results show its feasibility and effectiveness. Then the original distribution with multiple peaks is approximated where the difference between use of identical variance and use of variant variance for each peak is also discussed. Finally as two applications, the smoothing of histogram and evaluation of data quantity are introduced to demonstrate utilization of proposed method.
<ロボティクス>
  • 佐藤 重彦, 野澤 昭雄, 井出 英人
    2004 年 124 巻 7 号 p. 1390-1395
    発行日: 2004年
    公開日: 2004/10/01
    ジャーナル フリー
    Cooperated multi robots system has much dominance in comparison with single robot system. It is able to adapt to various circumstances and has a flexibility for variation of tasks. However it has still problems to control each robot, though methods for control multi robots system have been studied. Recently, the robots have been coming into real scene. And emotion and sensitivity of the robots have been widely studied.
    In this study, human emotion model based on psychological interaction was adapt to multi robots system to achieve methods for organization of multi robots. The characteristics of behavior of multi robots system achieved through computer simulation were analyzed. As a result, very complexed and interesting behavior was emerged even though it has rather simple configuration. And it has flexiblity in various circumstances. Additional experiment with actual robots will be conducted based on the emotion model.
<システム>
  • 伊藤 雅, 林田 巧
    2004 年 124 巻 7 号 p. 1396-1403
    発行日: 2004年
    公開日: 2004/10/01
    ジャーナル フリー
    This paper presents an optimization method for guitar fingering. The fingering is to determine a unique combination of string, fret and finger corresponding to the note. The method aims to generate the best fingering pattern for guitar robots rather than beginners. Furthermore, it can be applied to any musical score on single notes. A fingering action can be decomposed into three motions, that is, a motion of press string, release string and move fretting hand. The cost for moving the hand is estimated on the basis of Manhattan distance which is the sum of distances along fret and string directions. The objective is to minimize the total fingering costs, subject to fret, string and finger constraints. As a sequence of notes on the score forms a line on time series, the optimization for guitar fingering can be resolved into a multistage decision problem. Dynamic programming is exceedingly effective to solve such a problem. A level concept is introduced into rendering states so as to make multiple DP solutions lead a unique one among the DP backward processes. For example, if two fingerings have the same value of cost at different states on a stage, then the low position would be taken precedence over the high position, and the index finger would be over the middle finger.
<ソフトウェア・情報処理>
論文
<電子材料>
  • 柳田 英明, 上村 一秀, 横谷 篤至, 黒澤 宏
    2004 年 124 巻 7 号 p. 1410-1415
    発行日: 2004年
    公開日: 2004/10/01
    ジャーナル フリー
    Tetraethoxysilane (TEOS : SiO4(C2H5)4) is widely used to fabricate silicon-dioxide insulator thin films in LSI device technologies. We have already reported that TEOS is photo-dissociated to result in oxide layer deposition at room temperature by a vacuum ultraviolet excimer lamp (λ=172 nm). In this paper, we have observed the initial stages of oxide layer deposition on silicon clean surfaces with the scanning tunneling microscope (STM) and time-of-flight (TOF) mass spectrometer. An argon excimer lamp (λ=126 nm) is used for the photo-chemical vapor deposition (CVD). TEOS molecules have been dissociated to be SiO4(C2H5)n(n=1, 2, 3) groups on the clean surfaces. The 126 nm photons dissociate the molecules or radicals to Si-Om(m=1, 3, 4) after 2 min. Finally, the Si-Om adsorbed molecules and Si atoms in the substrate are re-arranged to be a disorder structure by the photons.
<電子・集積回路>
  • Kei Eguchi, Hongbing Zhu, Toru Tabata, Fumio Ueno, Takahiro Inoue
    2004 年 124 巻 7 号 p. 1416-1421
    発行日: 2004年
    公開日: 2004/10/01
    ジャーナル フリー
    In this paper, a Dickson-type power converter with bootstrapped gate transfer switches is proposed. The circuit is designed by adding a polarity exchange circuit and N (N=2, 3, ...) power-switches to the conventional Dickson-type circuit. These additional circuits enable the converter to provide not only a DC-DC output but a DC-AC output. In the process of DC-AC conversion, the circuit can generate various AC outputs by exploiting pulse amplitude modulation (PAM) method. Furthermore, high efficiency can be achieved by adopting novel bootstrap circuits to power-switches in the converter. Via maximum circuits constructed with diodes, the voltage of forward stage is charged to a capacitor in the bootstrap circuit. By connecting the charged-capacitor between the gate terminals of power-switches and the output terminal of the maximum circuit, the bootstrap circuits reduce the on-resistance of the power-switches. The SPICE simulations for the proposed circuit show the following results: 1. the voltage efficiency of DC-DC conversion is more than 90 % when the output load Ro = 500Ω and 2. various types of AC outputs can be generated.
  • 秋山 登, 菊地 睦, 佐瀬 隆志, 稲葉 政光, 渡辺 篤雄, 村林 文夫, 工藤 博之
    2004 年 124 巻 7 号 p. 1422-1427
    発行日: 2004年
    公開日: 2004/10/01
    ジャーナル フリー
    The demand for effective production control and flexible manufacturing lines is increasing placing many kinds of products on the market promptly according to various needs. Therefore field networks that connect line controllers and field equipment like sensors and actuators have been spreading instead of separate conventional electrical wiring. We describe the development of a one-chip transceiver LSI for communication interface. High-voltage devices with SOI substrate and monolithic isolator technology were developed. Then, a 1 MHz transceiver circuit and discrete devices, including 1 kV isolators, and a 5 V switching regulator were integrated onto the one-chip that measures 4.5 mm x 4.3 mm. An LSI with a monolithic isolator and a switching regulator can provide a small, low-consumption communication interface for the DeviceNet of a representative field network. Consequently, the mounting space and power consumption of the interface were reduced by 50% and 80%, respectively. We confirmed that an interface circuit employing the LSI in a network control system for air conditioners had the same transmission waveforms as that of a conventional interface circuit.
<医用電子・生体工学>
<音声画像処理・認識>
  • 木許 雅則, 古川 利博
    2004 年 124 巻 7 号 p. 1437-1445
    発行日: 2004年
    公開日: 2004/10/01
    ジャーナル フリー
    J. Benesty et al. indicated that a method of a direct generalization of the affine projection algorithm (APA) to a multichannel case converges very slowly because the cross-correlation between input signals to all channels was not considered in it. MC-APA has been proposed to resolve this drawback. However, MC-APA carries out the update of coefficients of one channel by utilizing the properties of the input signal of the other channels, but not that of all channels. To overcome the defect of MC-APA, we proposed a new method in which all channel input signals are utilized for the coefficient correction of one channel. We show by many numerical examples that the proposed method have good convergence characteristics.
  • 佐藤 信夫, 林 隆史, 森 俊二, 畑岡 信夫
    2004 年 124 巻 7 号 p. 1446-1454
    発行日: 2004年
    公開日: 2004/10/01
    ジャーナル フリー
    The pyramid hierarchical structure is known as a framework which is fit to extract global features from local ones through hierarchically arranged multi-resolution planes. However, there is a problem in which a mapping process between consecutive planes is complicated and not so effective in a line extraction.
    In this paper we propose a new method devising a new mapping between consecutive planes. We show that the proposed method is robust against noise and effective to extract and describe the line structures. We examine the system performance for noisy images and several kinds of line structures comparing with the Hough transformation method and confirm that the proposed method can extract and describe the line structures. Furthermore, we demonstrate that the vectorization of a property land map as real data can be done effectively with high precision and noise removing is performed automatically.
<制御・計測>
  • 玉置 守宙, 秋月 影雄, 大浦 邦彦
    2004 年 124 巻 7 号 p. 1455-1461
    発行日: 2004年
    公開日: 2004/10/01
    ジャーナル フリー
    This paper proposes an identification algorithm for time-varying system. We apply subspace method for estimation, since it is known as useful when the input-output (I/O) data are observed by multi-input multi-output (MIMO) system. Among many proposed techniques of subspace methods, we use MOESP (MIMO Output-Error State Space Model Identification) in this paper, which assures arithmetic stability by RQ-factorization and singular-value decomposition (SVD). Generally, subspace methods can be applied after I/O data collection, so that we introduce updated steps of matrices for PI-MOESP, which uses past inputs for instrumental variables. We propose recursive update algorithm of PI-MOESP, including estimation step of the system order, and consider some parameters inherent to the algorithm, those are initial number of data, estimation step of the order and forgetting factor. Numerical example shows usefulness of the proposed method.
  • 磯永 聡, 服部 哲, 岡本 昌幸, 田中 正吾
    2004 年 124 巻 7 号 p. 1462-1470
    発行日: 2004年
    公開日: 2004/10/01
    ジャーナル フリー
    Developing the monitoring system in a bathroom is important to prevent accidents in there. As a monitoring device, CCD camera is easily imagined. But this device is unsuitable by various problems such as violating one’s privacy and moistures on the lens, and so on. This paper proposes a new idea of immersing a pipe, which an acoustic sensor is set inside, in the water. By this procedure, we can obtain not only a direct sound information from the people in the bathroom but also the water level information. That is, stationary waves depending on the water level will be formed inside the pipe and the frequency analysis of the stationary waves enables the real time measurement of the level.
    The proposed measurement system adaptively selects the data window, under which the accurate measurement is automatically achieved, by applying the criterion of Fisher’s information matrix. The experiments will demonstrate the accurate water level measurement by the proposed system regardless of the sound amplitude and type.
  • 匹田 志朗, 岩田 雅史, 阿部 茂
    2004 年 124 巻 7 号 p. 1471-1477
    発行日: 2004年
    公開日: 2004/10/01
    ジャーナル フリー
    We propose a new elevator group control method with destination call entry system for up-peak periods such as the morning peak in office buildings. This method can reduce the passengers’ average waiting time, by allocating an appropriate number of destination calls to each car. Also the real-time simulation is used to determine the optimal number of calls. The results of some simulations are included to verify the effectiveness of the proposed method.
<ソフトコンピューティング>
  • 梅迫 公輔, 大林 正直, 小林 邦和
    2004 年 124 巻 7 号 p. 1478-1483
    発行日: 2004年
    公開日: 2004/10/01
    ジャーナル フリー
    In this paper, an evolutionary reinforcement learning system with time-varying parameters that can learn appropriate policy in dynamical POMDPs is proposed. The proposed system has time-varying parameters that can be adjusted by using reinforcement learning. Hence, the system can adapt to the time variation of the dynamical environment even if its variation cannot be observed. In addition, the state space of the environment is divided evolutionarily. So, one need not to divide the state space in advance. The efficacy of the proposed system is shown by mobile robot control simulation under the environment belongs to dynamical POMDPs. The environment is the passage that has gates iterate opening and closing.
<システム>
  • 安田 恵一郎, 金澤 貴彦
    2004 年 124 巻 7 号 p. 1484-1493
    発行日: 2004年
    公開日: 2004/10/01
    ジャーナル フリー
    Most of the actual optimization problems are large-scale, nonlinear and multi-peaked (nonconvex). Furthermore, with the widespread use of high-speed and large capacity computers as the background, it is felt necessary in recent years to derive a global solution for nonlinear and multi-peaked optimization problems. It is one of the most important topics in optimization. This paper proposes a new dynamic tunneling algorithm with multi trajectories (Multi-Trajectory Dynamic Tunneling Algorithm) that is composed of two systems, an optimization system and a tunneling system. The numerical stability of the conventional dynamic tunneling algorithm is theoretically investigated and interaction between each trajectory of the tunneling system is introduced in order to improve search efficiency. The proposed algorithm is applied to 2-variable and 10-variable typical multi-peaked nonlinear optimization problems.
  • Rong Long Wang, Zheng Tang, Xin Shun Xu
    2004 年 124 巻 7 号 p. 1494-1499
    発行日: 2004年
    公開日: 2004/10/01
    ジャーナル フリー
    An efficient parallel algorithm for solving the minimum vertex cover problem using binary neural network is presented. The proposed algorithm which is designed to find the smallest vertex cover of a graph, uses the binary neural network to get a near-smallest vertex cover of the graph, and adjusts the balance between the constraint term and the cost term of the energy function to help the network escape from the state of the near-smallest vertex cover to the state of the smallest vertex cover or better one. The proposed algorithm is tested on a large number of random graphs and benchmark graphs. The simulation results show that the proposed algorithm is very satisfactory and better than previous works for solving the minimum vertex cover problem.
<ソフトウェア・情報処理>
  • 柘植 覚, 獅々堀 正幹, 黒岩 眞吾, 北 研二
    2004 年 124 巻 7 号 p. 1500-1506
    発行日: 2004年
    公開日: 2004/10/01
    ジャーナル フリー
    The Vector Space Model (VSM) is a conventional information retrieval model, which represents a document collection by a term-by-document matrix. Since term-by-document matrices are usually high-dimensional and sparse, they are susceptible to noise and are also difficult to capture the underlying semantic structure. Dimensionality reduction is a way to overcome these problems. Principal Component Analysis (PCA) and Singular Value Decomposition (SVD) are popular techniques for dimensionality reduction based on matrix decomposition, however they contain both positive and negative values in the decomposed matrices. In the work described here, we use Non-negative Matrix Factorization (NMF) for dimensionality reduction of the vector space model. Since matrices decomposed by NMF only contain non-negative values, the original data are represented by only additive, not subtractive, combinations of the basis vectors. This characteristic of parts-based representation is appealing because it reflects the intuitive notion of combining parts to form a whole. Using MEDLINE collection, we experimentally showed that NMF offers great improvement over the vector space model.
  • 松田 善臣, 名嘉村 盛和, 姜 東植, 宮城 隼夫
    2004 年 124 巻 7 号 p. 1507-1514
    発行日: 2004年
    公開日: 2004/10/01
    ジャーナル フリー
    The paper treats an optimal routing problem for sightseeing (ORPS) in which we consider time varying travel time and location value. ORPS is defined on a complete graph in which location values are associated with node weight while travel time depends on edge weight. The aim of the problem is to construct a path with maximal total value under the condition that the total travel time does not exceed a time limit. In our model, the location value and the travel time are given as functions with respect to time. For the case that all the time functions are represented as piece-wise constant, we verify its ΝΡ-completeness.
    We present an algorithm to find the exact solution based on the backtrack searching and also propose a heuristic algorithm based on the time domain decomposition approach. The heuristic algorithm can lead good quality of approximate solutions, with drastically reduced computation time. To evaluate the algorithms, a computational experiment is given. The experimental results demonstrate that computing time by exact algorithm grow exponentially with the increase in nodes and time limits, whereas the heuristic approach derive approximate solution of 92% quality to the exact one within one second for practical problem size.
研究開発レター
<音声画像処理・認識>
エラータ
feedback
Top