The number of subscribers for broadband Internet services on ADSL and CATV is rapidly increasing in Japan. Furthermore the Japanese Government has e-Japan Program which sets the Japanese strategy to realize the world top-level Internet inflastructure in 2005. To meet the strong demands for high speed Internet access, the photonic network technologies are essential. This article provides brief introduction of three papers included in this issue.
The abrupt growth of the Internet has led to a telecommunications revolution whose range extends from voice services to multimedia and ubiquitous services. The new services require new high speed networks with a large capacity and system flexibility. Optical networks based on time division multiolexing (TDM) and wavelength division multipexing (WDM) technologies are a promising solution. This paper describes recent progress on hyper photonic devices, such as integrated semiconductor laser diodes, optical planar lightwave circuit devies and opticl fiber amplifiers, which are keys to development of optical network systems.
Chromatic dispersion in optical fibers is one of the most important parameters that determine the performance of transmission system. Especially, in the case of recent high-bit-rate and long-haul transmission systems, it is necessary to keep precisely track of the dispersion characteristics of fibers. And the total dispersion of system has to be close to zero by compensating the fiber dispersion, because the dispersion tolerance is very small in such high-bit-rate systems. Firstly in this document, the author explains the basics of chromatic dispersion in fibers and its characteristics of typical fibers. Inthe next, the principle of dispersion compensation techniques and their trends are introduced.
We propose an interferometric-type multi-state optical switch and an interferometric-type digital performance optical switch using fiber bragg grating (FBG) and the phase reversal effect. The effect is a reversing phenomenon of optical phase difference between the reflected and transmitted light from an FBG on a reflectivity spectrum dip between the main reflection band and first sideband. By combining this effect an reflectivitychange, the multi-state swich including only a single FBG can perform as a 4-state switch. In this paper we describe the principle of the phase reversal effect and some novel optical switch structures, and also the results of experimental examination.
Configurations of a Raman amplifier suitable for a 1.3-μm wavelength band are discussed and their properies are experimentally investigated. A pump light with a wavelength of 1.23-μm that is necessary for the Raman amplification in the 1.3-μm wavelength band is obtained using a 1.06-μm fiber laser and a Raman laser technique. Concerning the Raman laser, wavelength conversion from 1.06-μm to 1.23-μm is effectively achieved using a cavity configuration including fiber bragg gratings and a dispersion shifted fiber. On the other hand, a conventional dispersion compensation fiber which has an essential property of high non-linearity is applied in order to obtain large gain at 1.3-μm Net gain of 35 dB and ouput power of 15 dBm are achieved. To confirm the applicability of the Raman amplifiers to high-speed optical transmissions, experiments of 20 Gbit/s×2 WDM repeaterless transmission through a 80-km conventional single mode fiber is carried out. The 1.3-μm signal should be degraded due to the dispersion caused by the dispersion compensation fiber in the Raman amplifier. However, bit error rate of less than 10-12 is obtained at both wavelengths, which is enough performance for a practical use.
Characteristics of single channel 80-Gbit/s optical signal transmissions using a standard single mode fiber (SMF) are discussed including the effect of alternate polarization multiplexing at the sending end. Allowable values of transmission line length, input optical powers to th SMF, and dispersion margin are evaluated based on bit error rate measurements for three line configurations, unrepeated lines, lines having a repeater with a 100-km spacing, and multiply repeated lines with a 60-km spacing. As for the transmission line length, error free transmissions are experimentally possible for the total lengths of 160 km, 200 km, and 300 km for the individal cases. Unrepeated 160-km transmission experiments where an 80-Gbit/s optical signal with a single polarization is launched in the SMF are made for comparison. Superiority of the alternate polarization multiplexing is confirmed through the measuremets of input optical power, optical pulse amplitude fluctuation, and optical spectrum broadening.
Waveform distortion of logic signal, by accelerating clock speed and improving implementation density, makes non-linear output for the input and has been giving behavioral problem to electronic system. There are overshoot type, hysteresis type, and undershoot type for transient waveform distortion. Deep undershoot could cause latchup for CMOS logic IC, data destruction for DRAM, pulse split for TTL IC, respectively. Undershoot of logic signal is not discussed only by transient phenomena (time domain ) but also by filter theory ( frequency domain). Time sequential resistance R variation by output circuit impedance variation causes such time domain variations as damping constant α and damping coefficient ζ. While resistance R variation causes Q variation in frequency domain. In this paper, an occurrence mechanism is shown that undershoot occurs when parasitic series resonance circuits are composed at load side of driving device and damping coefficient ζ varies time sequentially for totempole or complimentary circuits. PSpice simulation suggests that undershoot is reduced by controlling the phase of punch through current.
A step-up and step-down DC-DC converter with cell structure is proposed in this paper. The proposed circuit is designed by using switched-capacitor (SC) techniques. The cell structure of this circuit enables us to synthesize the DC-DC converter with small number of power-switches. Furthermore, it can provide small and simple structure of circuits to generate clock-pulses. Concerning the proposed DC-DC converter, SPICE simulations are performed to investigate the characteristics of the circuit. The SPICE simulations show that 1. the number of power switches for the proposed circuit is less than that for the previous circuit, 2. the settling time is less than 6μs, and 3. the efficiency is more than 90%. Furthermore, the validity of circuit design is confirmed by experiments. The proposed DC-DC converter will be useful as a building block of mobile equipments.
In the human visual system, the visual stimuli are first received and preprocessed in the retina and then transmitted to the cerebral visual cortex via the lateral geniculate body. It is reported by Campbell and his associates that the cerebral visual cortex conducts processing consisting of spatial-frequency-and-orientation-specific multiple narrow-band channels. It is also reported that spatial-frequency-and-orientation-specific visual stimuli cause adaptations in the corresponding channels, resulting in the reductions of visual sensitivity around the frequency range. In this article, we focused our attention on the multiple channel theory that Campbell et al. proposed, and constructed a model of the human visual system that had more than one frequency channel. We have conducted various simulation studies using the above model. The results are compared with psychophysical data. In the course of the simulation studies using Landolt rings, to measure visual acuity, we happened to find spatial-frequency-specific adaptations will cause paradoxical transient improvements in visual acuity. Therefore we conducted perceptual experiments to see if the improvements in visual acuity take place after spatial-frequency-specific visual adaptation. As a whole, paradoxical transient improvements in visual acuity seem to take place.
Monosynaptic propagation delays of action potentials between neurons are the important parameters in the study of neural network dynamics. This paper proposes a new method named WFDEM, Window-Free Delay-Estimation Method, to estimate the propagation delays from neuronal spike trains recorded extracellularly. WFDEM is derived as a maximum likelihood estimation method, and can be conducted in a reasonable time by using EM algorithm after some approximations. The other advantages of the new method include (1) it does not need to set appropriate time bins, and (2) it can possibly make use of the advantage of the simultaneous multi-electrode recording. Simulation study shows that WFDEM can achieve a high estimation performance when it is initialized by rough delay estimations that can be obtained by CCH (cross-coincidence histogram), especially when many pre-synaptic neurons are recorded simultaneously.
In order to detect ischemic and/or infarcted myocardium of a left ventricle, a motion of the cardiac muscle gives important information. Conventional method, such as center line method, may observe and measure the only overall motion of the myocardium using echocardiography. In this article, we propose a new method to estimate using block matching. Then, applying this method to the echocardiograph, we estimate not only the overall motion but also local displacement of the myocardium on the echocardiogram.
Many yellow colored markers are placed on pavement and in public facilities for sight handicapped persons. If a small electronics mounted in a white cane able to detect the yellow marker, it will be more useful for cane's owner. In this paper, we propose a realization such a small electronics. Our effort has been devoted to ensure for fair detection even in dark and lessen its wight and power consumption.
The purpose of this study is to apply the sensing mechanism of plants to a new sensing system in the field of engineering. In this paper, it is presented that the sensing equipment which imitates the function of phytochrome in plants makes structure conversion, PR (red light receptor) and PFR (far-red light receptor), and its germination control. The sensor systems of this equipment have characteristics of control for switching on-off in accordance with the condition of environment. The performance of calculation circuits of this equipment was checked. As a result, it was clarified that this sensing equipment imitating the functions of phototransformation and germination control of plants was sufficiently operating.
This paper presents a damage estimation model of an earthquake using interferometric SAR data pairs. Coherence is defined as a correlation function of two single look complex (SLC) SAR data. The coherence derived from multi-temporal SLCs before and after the earthquake includes a temporal decorrelation in disturbed regions. To estimate the damage degree using the temporal decorrelation, the proposed model employs a coherence ratio which is the ratio of the coherence derived from SLCs before and after the event to the coherence derived from a pair of SLCs before the event. The temporal coherence ratio is computed by compensating the coherence ratio with the spacial decorrelation. Damage degrees are defined by logarithmic probability range of damaged area ratio. We indicate the significant relationship between the temporal coherence ratio and the probability of the damage degree. In the case of the Hyogoken-Nanbu Earthquake, its relationship can be represented by a linear function. Applying the least squares, we determine the parameters of regression lines and have succeeded in estimating the cumulative probability of the damage degree from the temporal coherence ratio.
This paper discusses face and hand-sign recognition and its applications to a multi-camera system. We propose a vision-based interface that focuses on faces and hand for recognition. Using a stereo camera system, face and hand regions are detected by the referenced skin color method, and the distances of the targets are then estimated. The positions and sizes of targets are evaluated for tracking between consecutive frames. Two tracking cameras perform tracking of such targets as faces and hands, and also zoom in on the selected target to acquire an appropriately sized image for recognition. We experimented with a hierarchical linear discriminant analysis that distinguishes between faces and hands on the first step, and then recognizes face and hand-signs in each discriminant space by using four directional features. These experiments showed that our system could distinguish between faces and hands, as well as recognize face and hand-signs.
We have proposed a system for extracting bibliography information from still color image of a book cover. Although it works efficiently for ordinary books, it seems impossible to obtain satisfying results with the exact same system for the book covers containing gilt characters. The reason is that while photographing a book cover onto the image plane, reflection effects on the book cover which contains gilt characters are much more complicated than that on ordinary ones, so that it is impossible to obtain a still image satisfied enough to be processed by the system. When photographing a book cover into a still image, for ordinary book covers there usually only is total diffuse reflection effects. While for those contain gilt characters, there is mixed reflection effects-diffuse reflection and mirror reflection. That mixed effects make some part of book cover invisible in the still image. Instead of a still image, image sequence photographed by a moving camera will be helpful. The reflection on one spatial point is changing because the effects of mirror reflection in different frames. That means that the intensity of the pixel corresponding to that spatial point could be bright enough to be visible in some frames and could be too dark to be visible in others. This paper describes an efficient method for extracting characters from color image sequence of book covers containing gilt characters, taking advantage of reflection changes in different frames to extract enough reflection information for proper character extraction. Integrated with this new method, the proposed system is expected to work more accurately.
A lineament is one of the most important features showing subsurface elements or structural weakness such as faults and is usually extracted by visual analysis of enhanced image data. A different expert may extract different segments through a visual approach. To detect lineaments under the same conditions, this paper proposes a data processing approach on the automatically extraction of edge from Landsat Thematic Mapper (TM) data. The method can delete the unnecessary edge and decide the threshold for the binary processing considering the features of the histogram of the data. Relief energy maps help visually understand the circumstances for incursions of rivers. The relief energy has been computed, and then the correspondence between the computation and extracted edges is described. As a result of experience, it is evident that edges from each band TM data exist in the region where the relief energy changes and its saddle. It becomes also clear that the edges obtained from each band TM data with the proposed method are in agreement with lineaments by experts through the relief energy map.
Automatic quality recognition of interphones has been well developed and it is important that the classifier must be highly accurate. Generally, the accuracy is represented by a recognition rate for sample data. To evaluate it more strictly, we will introduce a reliability criterion. In the pattern recognition, neural networks (NNs) have been adopted. Among them a competitive NN has a simple structure and can explain the input/output relation more easily than a layered NN based on the error back-propagation method. Thus, we adopt a competitive NN for the recognition of good and no good interphones where the learning vector quantization (LVQ) method is used for training. Finally, we show the effectiveness of the proposed method by applying it to real interphone data.
The template matching is known as a general-purposed image processing method. However, the target image area to be searched must have the right position and the same scale as the template image. However the method to search the target image area that has a free location, inclination and scale in an objective image by using the genetic algorithm was proposed, the method can extract only a single target image area. This paper proposes a method to extract plural target image areas that are similar to the template image. The immune system that can search plural local solutions is applied to the conventional genetic algorithm. The experimental results show that the proper plural target image areas with free locations, inclinations and scales were extracted successfully by the proposed method because the method can search plural local solutions by keeping varieties of individuals.
In recent years, research studies and application of face recognition have become important, as face recognition is widely used in a variety of fields such as security systems, databases and so on. Herein we propose a method for detecting each facial region from any given frame in a moving picture. Moving pictures unavoidably include incidences of occlusion facial regions; our method can extract distinct facial regions in mobile situations, which include occlusion facial region. To detect facial regions, we utilized color and shape input and movement vectors. Computer simulation of the proposed method is given showing the effectiveness in occlusion facial regions, and extraction rate of facial region resulted 93.3% during 1200 frames.
We propose a learning model of sound source localization in a non-linear motor system with self-organizing through the interaction between motion and audio-visual sensing. This ability is acquired by repeated sensing and moving. The model consists of two modules -a visual estimation module and an auditory estimation module. These modules consist of a self-organizing feature map. These two modules learn at once. The visual estimation module learns by direct inverse modeling and the auditory estimation module uses the output from the visual estimation module to learn. We conducted computer simulation experiments to investigate the validity of the proposed module. The experimental results demonstrate that sound source localization ability can be acquired without supervision for a nonlinear system and the system is robust for noise and experimental environment.
This paper presents a gradient ascent learning algorithm in weight domain for solving local minimum problem of a Hopfield network. The learning algorithm has two phases, the time domain phase and the weight domain phase. The former seeks a minimum of the energy function by updating states in time domain of a Hopfield network, and the latter intentionally increases the energy of the network by modifying weights in a gradient ascent direction of the energy in weight domain after the network update stabilization in time domain is achieved. The two phases are repeated until a condition, for example the energy function E=0 is satisfied. The learning algorithm is applied to a two-neuron Hopfield network and an N-queen problem, extensive simulations are performed and its effectiveness is confirmed.
As the deregulation of electric business proceeds, each company needs to construct a risk hedging system, which the company has not been taking much care so far. In the electric business, the system risk should be treated more carefully because it may bring terrible financial loss to a generation company. In this paper, we address the system risk hedging issue. First, we introduce definitions for the system risk, then we discuss two typical hedging methodologies, insurance and catastrophe option. Finally, we suggest the optimal strategy, which includes risk-controlling operation.
In this paper, we investigated the new passive ranging system to located the moving target. Then we computed the time-delay difference on the output of a cross-correlater between two radiated noise signals received by the acoustic sensors, and we solved three-dimensional equations for the acoustic velocity and these sensor's co-ordinates positions. Still more, we investigated the issue point for the new passive ranging system are applied to the radiated noise of the vessel. Therefore, we analised the radiated noise signal of the target fishing boat which were collected in the test sea zone for clearing the acoustic characteristic, sampling the trackingsignal and computing the time-delay difference between two radiated noise signals received by the acoustic sensors.
Up to this time, there are two kinds of the passive ranging system. One thing is the seabed mounted transponders method and another is the float buoy ranging method. Both methods use the vessel-mounted pinger system. But when we need to estimate the running ability of the underwater vehicle easily and economically, we can't use the expensive vessel-mounted pinger system. So we must tract down the running locus of the underwater vehicle from the radiated noise of the vehicle. In this study, we investigated the new passive ranging system which located the moving target. Then we computed the time-delay difference based on the output of a cross-correlter between two radiated noise signals received by the acoustic sensors and we solved three-dimensional equations for the acoustic velocity and these sensor's coordinates positions to tract down the running locus of the underwater vehicle. In this paper, we verified an algorithm for the new passive ranging system by a simulation analysis, in which the radiated noise, the velocity and advance direction of the underwater vehicle are assumed. The simulation analysis cleared the effectiveness of an algorithm of the new passive ranging system to determine the accurate position of the underwater vehicle.
As microprocessors have been widely adopted and are integrated into various devices, the reliability of microprocessors themselves or micro-computer systems using such microprocessors is becoming a key element. With regard to the microcomputer systems, a highly promising way for improving their reliability will be to construct a fault tolerant system by using a processor module that has more than one microprocessors of the same function. In this paper, we will describe an architecture of a fault tolerant microprocessor and how to realize this. By using this microprocessor we can construct a practical fault tolerant system without any influences to both hardware and software. This microprocessor is a combination of a core processor and a fault tolerant supporting module. The core processor is based on the general-purpose 32-bit microprocessor, GMICRO/100. This microprocessor executes the following fault tolerance operations automatically detection of fault in the processor, rollback by means of hardware for the recovery from transient failure, self-diagnosis for identifying a fault processor and system reconfiguration by the processor on the normal side for the recovery from permanent failures.
As Internet technology is widely spread, B to B Electronic Commerce (EC) is expanding into practical business field. Market mechanism based on general equilibrium theory in microeconomics solves product distribution problem by allocating the scheduled resources according to market prices. We propose a B to B EC server including virtual market with the market-oriented programming that mediates amongst unspecified various companies in the trade, and demonstrate the applicability of the economic analysis to this framework. The proposed server facilitates sophisticated B to B EC, which conducts a Pareto optimal solution for all the participating business units.
A clock recovery based on optical modulation and an electric phase lock loop is demonstrated for 80-Gbit/s optical signals. The 80-Gbit/s pulses are modulated by an optical modulator with a frequency of 40.4GHz and then converted into electric signals. The phase of the 39.6-GHz signal component included in the converted signal is compared with that fed from a voltage-controlled oscillator. The scheme is successfully applied to a 400-km transmission experiment.
Turbo codes, also known as parallel concatenated convolutional codes (PCCC), were demonstrated to have performance within 0.5 dB of the Shannon limit around BER=10-5 over AWGN channel. However, these simulation results were only based on large interleaver size and many decoding iterations. In personal communication systems, the standard rate of tone is 9.6 kbit/s and the time interval of each frame is 20ms. So the length of a transmission frame is about 192 bits and is much shorter than the 256×256 frame. When the short frame transmission is taken into consideration, a proper interleaver design is very important. In this letter, the theoretical analysis of the performance of an interleaver based on the correlation coefficient is presented. An “inverse diagonal” interleaver is proposed to show that it has the lower correlation coefficient compared with the conventional row-column block interleaver design and is easily realized.