IEEJ Transactions on Electronics, Information and Systems
Online ISSN : 1348-8155
Print ISSN : 0385-4221
ISSN-L : 0385-4221
Volume 110, Issue 9
Displaying 1-8 of 8 articles from this issue
  • Yoichi Fujii
    1990Volume 110Issue 9 Pages 517-520
    Published: September 20, 1990
    Released on J-STAGE: December 19, 2008
    JOURNAL FREE ACCESS
    Download PDF (816K)
  • Kenji Kanayama, Takeshi Naitoh
    1990Volume 110Issue 9 Pages 521-528
    Published: September 20, 1990
    Released on J-STAGE: December 19, 2008
    JOURNAL FREE ACCESS
    At the paper producing line, it is required to judge whether the defect which has occurred on producing process, shall be rejected or not at the next stage. In this paper we propose an automatic defect picture recording system, in which our FT-CCD shutter video camera with the external trigger is used. This system has been developed for the purpose of increasing the efficiency of paper production.
    First of all, we describe the system architecture and software methods of (1) Real-time multichannel trigger control operation, (2) Serial defect recording operation, (3) Defect image data file management operation. And we found that a wide-type reflector lamp is suitable for flat lighting. In a practical field test, we have confirmed that this system is capable of monitoring a 0.5mm-diameter defect even if the paper sheet moves at the high speed of 800m/min. This system is expected to be used for not only paper but also for various kinds of sheet such as aluminum, film, steel, and others, as the facility for high speed and high accuracy inspection at producing lines.
    Download PDF (1501K)
  • Hideya Takahashi, Kenji Matsushita, Eiji Shimizu
    1990Volume 110Issue 9 Pages 529-533
    Published: September 20, 1990
    Released on J-STAGE: December 19, 2008
    JOURNAL FREE ACCESS
    The innate parallelism, noninter fering transmittance, and wide bandwidth of optics have possibilities of both large capacity and high speed in computation. However, the optical computation based on the binary numbers has not taken full advantage of the intrinsic parallelism of optics because of carries. On the other hand, since the modefied signed-digit (MSD) number system offers carry-free addition and subtraction, optical implementation of the MSD arithmetic can take full advantage of intrinsic parallelism of optics. But, the set of this number system includes a negative number, so the optical implementation using optical information is not suitable for the MSD arithmetic. We propose the optical MSD computation system by optical addition and subtraction using optical Fourier transform which uses optical phase information.
    Download PDF (1130K)
  • Masanari Taniguchi, Tasuku Takagi
    1990Volume 110Issue 9 Pages 534-539
    Published: September 20, 1990
    Released on J-STAGE: December 19, 2008
    JOURNAL FREE ACCESS
    A new measuring system (Holographic Pattern Measuring System, HPMS) has been developed to the thermal deformation analysis of printed circuit board (PCB) due to thermal stress. The HPMS combines both techniques of holography and graphic image processing. The tempaeature on the PCB surface was measured by using a thermographic video system, and the relationship between temperature and displacement of the PCB was studied. As a result, the microscopical displacement of the displacement can be quantitatively measured, and the distribution of the displacement can be displayed as a three dimensional graphic image. The temperature distribution of the PCB surface showed the various relationship between the characteristics of the deformation. Therefore, the holographic pattern testing method may be useful in analysis of the detail thermal stress evaluation of the PCB.
    Download PDF (1799K)
  • Shoshiro Hatakeyama, Masayuki Takahashi, Kazuyuki Aihara, Hisaaki Hata ...
    1990Volume 110Issue 9 Pages 540-546
    Published: September 20, 1990
    Released on J-STAGE: December 19, 2008
    JOURNAL FREE ACCESS
    Overall, the structural simplicity of the network has a strong relationship to its ability to promote generalization. With regards to this type of situation. Rumelhart proposed a new learning algorithm which has a common algorithm for backpropagation (BP) learning and an additional penalizing term, relating to the network complexity, which, when working in tandem, increase the ability of the back propagation learning procedure. This learning algorithm has a property that modifies the weighting factors, thereby, decreasing the total sum of both the errors and the complexity of the network simultaneously.
    The authors provide insight as to the implications of the above modified BP algorithm and furnish several simulations which test its performance. Moreover, a refining procedure is provided, this is based on the investigations into the balance between the error term and the complexity term. Finally, the refined algorithm is successfully applied to the systems identification problem and subsequently, shows the desired performance for generalization in neural networks.
    Download PDF (1270K)
  • Ken'iti Kido, Jiaxiang Liu, Hiroyuki Noto, Masato Abe
    1990Volume 110Issue 9 Pages 547-553
    Published: September 20, 1990
    Released on J-STAGE: December 19, 2008
    JOURNAL FREE ACCESS
    This paper proposes a method to locate sound sources and to estimate their spectra in a noisy environment using many sensors. It is found from some computer simulations and experiments that image sources appear if the number of sensors is insufficient. According to the investigation, it is found that not only the averaging in space domain by changing the combination of sensors but also the averaging in frequency domain is necessary for diminishing the image sources. The effects of diffraction in the space, the boundary wall, the flow of the medium etc. remaine for the further investigation for the practical use.
    Download PDF (1122K)
  • Toshikatsu Sonoda, Ryuzo Ueda
    1990Volume 110Issue 9 Pages 554-559
    Published: September 20, 1990
    Released on J-STAGE: December 19, 2008
    JOURNAL FREE ACCESS
    This paper presents a new magnetic field sensor of core type. The resolving power arrives at the order of 10-13T/√Hz at the room temperature. The sensor is composed of two bar type cores of which the B-H loop is desirable to have a high rectangularity ratio.
    The detection principle is analogous to the operation on a kind of equilibrium condition in ordinary bridge circuit. Each core is previously DC-biased in the opposite direction regarding the unknown magnetic field to be detected. A high frequency excitation field is superimposed onto each of DC field in anti-phase as an auxiliary signal and resultant search coil voltage induced by the auxiliary signal is null when the unknown field is zero. When the field is impressed to each core, the equilibrium condition is broken out and the voltage appears in proportion to it. Thus, we can detect the unknown field. For reduction of the noise in electronic circuit itself, some numbers of high frequency amplifier of the same characteristics are connected in parallel and a tuning amplifier is added to the output side of these amplifiers.
    By adopting the above approach, such the maximum possible resolving power can be achieved.
    Download PDF (2525K)
  • Kazuhisa Matsuo, Takehiko Tomikawa
    1990Volume 110Issue 9 Pages 560-561
    Published: September 20, 1990
    Released on J-STAGE: December 19, 2008
    JOURNAL FREE ACCESS
    Download PDF (681K)
feedback
Top