β-FeSi₂ thin films were grown on Si(111) substrates by pulsed laser deposition (PLD) method using Nd: YAG Laser (λ = 1064 nm, laser energy = 50 mJ, laser energy density = 1.65 J/cm², repetition frequency = 10 Hz). In the fabrication process, three targets of (a) Fe(5N), (b) FeSi₂(3N) and (c) Fe(5N) + FeSi₂(3N) were used. The β-FeSi₂ thin films having best properties of crystallinity were obtained in the case of (c), in which the first layer as the template was formed with the target of Fe(5N) and then, on top of that, the second layer was deposited with the target of FeSi₂(3N). At this time, it was found that by XRD measurement, the degree of crystallinity of the films of the case (c) in which the first layer was introduced as the template improved 1.4 times as much as compared with the case (a), and that by SEM and AFM observations, surface morphologies also improved. Moreover, it was found that by TEM obeservation, the β-FeSi₂ thin grew uniformly along the direction of (220) or (202) from the interface of the Si substrate and the film to the free surface and that by EDS analysis, the compositions of Fe and Si were uniformly distributed.

抄録全体を表示

Genetic Programming (GP) is an extension of Genetic Algorithm(GA) to handle more structural problems. In this paper, an approach to filter circuit design by GP is proposed. By designing a gene which includes not only the parameters of consisting elements, but also the structural information of the circuit, it becomes possible to apply the proposed approach to various types of filter circuits. GP depends much on trial and error due to its probabilitic nature. To decrease this uncertainty and ensure the progress of the evolution, Parallel GP with multiple populations with the island model is also proposed. An MPI-based cluster system is used for realization of this parallel computing where each island correspondsd to each node. A lowpass and an asymmetric bandpass filters are designed. One hundred times of trials for multiple populations with and without migrations are tested in the design of lowpass filter to confirm the validity of the proposed method. In the asymmetric bandpass filter design, the results are compared with those of the circuit designed by hand to confirm the effectiveness of the proposed method. The proposed approach is applicable to various types of filter circuits. It can contribute to an automated design procedure, where it would require a expirenced designer if done by hand. It is also possible to obtain a new circuit design which would not be possible if done by hand.

抄録全体を表示

For low-invasive laser treatment, stable absorption properties of tissue during laser treatment are essentially needed, which enable us to predict irradiation effect (ablation depth, thermally-altered depth) using them. There are water and proteins as main absorber in the mid-infrared range. During laser treatment, the absorption properties of soft tissue are changed due to laser-induced temperature rise and leading to the induction of unexpected irradiation effects. We have proposed and demonstrated a novel, low-invasive laser surgery with 6.1-μm-light irradiation. 6.1-μm-light can resonantly excite both OH bending (water) and amide-I (proteins). The absorption properties of soft tissue during laser treatment would be come considerably stable because they do not remarkably depend on the surrounding temperature and the water concentration. We examined the gelatin ablation properties by λ=6.1-μm irradiation using a tunable mid-infrared Free-Electron Laser and showed that λ=6.1-μm irradiation is one of the candidates for a low-invasive soft tissue incision. Gelatins were used as a model substrate for soft tissue. From these experiments, it was found as follows. (1) Enhanced gelatin ablation occurred near λ=6.1-μm, but not λ=6.45-μm (Amide II of proteins). (2) The gelatins were removed due to water vaporization, which is reflected in the static absorption properties of the gelatin before laser irradiation. Thus, we showed the applicability of λ=6.1-μm -light for a new light source to realize a low-invasive soft tissue incision.

抄録全体を表示

This paper proposes a using probe for measuring of contact force between tongue and palatal, during speech. We developed a 0.03 mm-thick stainless steel tongue force probe with a 3x5 mm force sensor at the tip. Linguopalatal contact force was measured by inserting the probe into the oral cavity. Contact force was measured at the following three locations. Based on the coordinate and measurement obtained at the three points, the action point of tongue force was calculated by the weighted mean. Linguopalatal contact force was measured in four adult men and women without articulation disorder and in three adult men with articulation disorders. Results showed that the action point of tongue force in subjects with articulation disorders was further toward the pharynx than that in subjects without articulation disorders. Linguopalatal contact pressure was then measured again by asking the subjects with articulation disorders to wear a palatal augmentation prosthesis (PAP) to compensate for insufficient linguopalatal contact force. The action point of tongue force became better approximated to that of subjects without articulation disorders. Given these results, our simple method for measuring linguopalatal contact force using a tongue force probe appears to be a promising tool for speech therapists treating articulation disorders.

抄録全体を表示

Barrier discharge characteristics have been investigated for twin needles-plane electrode configuration in dry air. Characteristics of partial discharge depending on the distance between the tips of twin needles have been focused. In the case of d equal to 1.0mm˜2.0mm (d : distance between the tips of twin needle), the discharge inception voltage value increased compared with that of the single needle-plane electrode configuration. Moreover, for the d equal to 4.0mm, the discharge inception voltage value became almost the same value in the case of the single needle. It is considered that the differences of the discharge inception voltage value have been caused by the influence of tip of needle each other. In addition, discharge images by intensified still camera and CCD camera, power consumption, and ozone concentration were examined.

抄録全体を表示

To get estimated traveling time and to analyze the traffic flow precisely, it is necessary to survey the traffic flow characteristics on each road. The traffic flow characteristics are represented by some factors such as speed and the number of vehicles passing through the road. In this paper, it proposed to measure the vehicle speed by the brightness analysis of video images.
Since a blur on the image analyzed by the frequency analysis of the image, it is considered that the vehicle speed is calculated by analyzing a blur image.
Simple Fourier transformation shows the periodic feature of the brightness noise as well as the vehicle length component and the blur component that includes the information of the vehicle speed. The periodic feature of the brightness noise prevents us from finding the periodic feature of the component of the vehicle speed. To exclude the unnecessary components from the analysis, two filters introduced into the brightness domain and the frequency domain. These filters make it possible to separate the vehicle speed component and the vehicle length component from the unnecessary components clearly. Then the vehicle speed component is calculated more precisely.

抄録全体を表示

The simultaneous perturbation stochastic approximation (SPSA) is an extension of Kiefer-Wolfowitz stochastic approximation algorithm. In SPSA, since all parameters are perturbed simultaneously, it is possible to modify parameters with only two measurements of an evaluation function regardless of the dimension of the parameter. We propose a parameter estimation algorithm using the SPSA. Convergence theorem for the proposed algorithm is shown. A simulation result also reveals a feasibility of the identification scheme proposed here.

抄録全体を表示

The trade-off between hardware and software is an important matter in hardware/software co-design. In embedded systems which develop immense software, the implement ability of software re-use serves as a key of development success. In this paper, we propose a co-design method using real-time evaluation of re-use parts. The co-design method uses processing time, memory, and number of gates as trade-off conditions. System level language SpecC is used for realizing the proposed method. We applied the method to ITS (Intelligent Transport Systems) safe driving support system and evaluated the effectiveness of the method.

抄録全体を表示

This paper presents a method for extracting profile information in tabular formats based on existing technologies called named entity extraction and information integration. Named entity extraction enables us to provide elements of tables for profile information. Information integration allows us to unify tables for making the profile information fruitful, though it requires predetermined initial tables. In this paper, we propose a whole system of extracting profile information by bridging the gap between the two technologies. For this purpose we employ a method of grouping named entities for making initial tables. For the extraction and grouping of named entities, we utilize support vector machines. Initial tables are then integrated if these are with the same name. From the experimental results on 7085 newspaper articles, we obtained the results of 53.8% precision with 58.7% recall; Although the proposed method is insufficient as a fully automated information extraction, it provides us a good starting point for extracting profile information.

抄録全体を表示

We have analyzed the drilling process with femtosecond laser on the silicon surface in order to investigate a degree of thermal effect during the dicing process of the very thin silicon substrate. A regenerative amplified Ti:Al₂O₃ laser (E= 30˜500 μJ/pulse, τ= 200 fs, λ= 780 nm, f= 10 Hz) was used and focused onto a 50 μm-thick silicon sample. ICCD (Intensified Charge coupled Device) camera with a high-speed gate of 5 ns was utilized to take images of processing hole. First, we investigated the dependence of laser energy on the speed of the formation of the drilled hole. As a result, it was found that the lager the energy, the slower the speed of the formation under the minimum hole was obtained. Consequently, in the case of defocused condition, even when the smaller the energy density was used, the very slow speed of formation and the much lager thermal effects are simultaneously observed. So we can say that the degree of the thermal effects is not simply related to energy density of the laser but strongly related to the speed of the formation, which can be measured by the ICCD camera. The similar tendency was also obtained for other materials, which are important for the fabrication of ICs (Al, Cu, SiO₂ and acrylic resin).

抄録全体を表示

A sensing system using plastic optical fibers and reagent papers was developed for the detection of urine and abnormal level of its components simultaneously. Among several components of urine the detection of two main components namely, protein and glucose was confirmed experimentally. Three states of the papers namely dry and wet with and without change in color, were taken into consideration. These three states were divided by setting the lower and upper threshold voltages at 2.2 V and 5.5 V, respectively. This system is considered to be simple in construction, easy to operate and cost-efficient.

抄録全体を表示

The low cost improved recognition of an uncooled infrared image processing has been required urgently for the development of high sensitive uncooled sensor like a cooled sensor. In this study, the uncooled infrared image taken by YBaCuO bolometer camera was improved with applying a special low cost image processing systems. The improved image with this processing exhibited not only the edge enhancement in a large contrast area but also the suppression of heat noise in a small contrast area. It is clarified that this improved recognition image processing is easy to realize at low cost, and the demand for this image processing will be expected to grow up as the use of uncooled infrared sensor system increases.

抄録全体を表示

In order to ride on the strength of paralell operation a feature of neural network, it is preferable that all neuron is implemented on hardware. Formerly, we combine Neural Network and ΔΣ modulation, which is a method of converting to 1bit pulsed signal. Then we succeeded to configurate “a Pulsed Neural Network based on ΔΣ modulation(DSM-PNN)", which keep the circuit scale as same as to operate precisely. In last paper, we proposed hardware implementation methods of DSM-PNN with GHA learning rule and show its availability in linear operation. However, since neural networks are characterized by nonlinear map, signals needs to be treated with sufficient precision, also in nonlinear operation. In this paper, in order to shows that the 1-bit signal processing by DSM-PNN can be available, even when it includes nonlinear operation, we proposed the technique of realizing algorithm of ICA including nonlinear operation in DSM-PNN and confirm the performance of it.

抄録全体を表示

Fundamental theory for minimizing the network cost per subscriber is investigated for broadband local loops. Long term analyses of broadband local loop cost per subscriber are overviewed to show the importance of minimizing cable installation costs.
First, the reduction of network installation cost is investigated in terms of network sharing. Sharing of a network among a broadband information network and a walk-assistance network is considered as an example. It is shown that network installation cost can be halved by network sharing.
Next, possibilities of further cost reduction is discussed. It is shown that the cost of walk-assistance network can be reduced by a factor of two by applying passive areas.
Finally, further reduction of network installation cost is investigated in terms of network sharing and cable-wireless combination. Dependence of network costs and passive area ratio on wireless area radius r is clarified. It is shown that the cost of a walk-assistance network can be reduced by a factor of four.

抄録全体を表示

Matching of stereo images is a fundamental task for 3D recovery. We have matched with feature points of a direct image and its mirror image by using GA and the information of apexes connections. However, when the feature points become large, miss matching were occurred because of the graph imbedding problem and the ambiguities of matching with occluded points. We introduce an immune algorithm (IA) to cope with these multi-peaks searching.
In this paper, we propose a matching algorithm of feature points/segments by using cut set graph and IA. First, the stereo images are specified to a criterion and a reference image, respectively. Both the feature points/segments of the images are classified to 3 groups; cut set, inner and outside group. An evaluation of reference image is defined how many the remark points/segments are included in each groups correctly correspond to these of the criterion groups, i. e, the evaluation is the degree of the fitness. The candidates of matching obtained from IA are selected by checking both the graphs connection conditions. We show that the peaks are effectively detected and matched correctly.

抄録全体を表示

This paper presents a speech enhancement using the classification between the dominants of speech and noise. In our system, a new classification scheme between the dominants of speech and noise is proposed. The proposed classifications use the standard deviation of the spectrum of observation signal in each band. We introduce two oversubtraction factors for the dominants of speech and noise, respectively. And spectral subtraction is carried out after the classification. The proposed method is tested on several noise types from the Noisex-92 database. From the investigation of segmental SNR, Itakura-Saito distance measure, inspection of spectrograms and listening tests, the proposed system is shown to be effective to reduce background noise. Moreover, the enhanced speech using our system generates less musical noise and distortion than that of conventional systems.

抄録全体を表示

This paper proposed a method to search a target image area that has a free location and inclination in an objective image. The gray-level histogram and the gray-level ratio of pixels pair histogram are used as the invariant feature to change of inclination in this method. The proposed method separates a circular template to ring regions with the common center and generates histograms about each separated ring region. These generated histograms are used for matching. Moreover, in order to reduce the computation cost with maintaining the reliability, effective ring regions for matching are selected by the genetic algorithm. The experimental results show that the target image areas were searched successfully in spite of some ring regions were not used and the high-speed matching was realized by using the generated concentric ring template image in comparison with by using the whole template image.

抄録全体を表示

A sensorless driving method of brushless DC Motors (BLDCM) using neural network has been studied in this paper. Considering the nonlinear characteristics and the parameter error of the modeling, neural networks are introduced to estimate the electromotive force (EMF). The results of simulation and experiment using offline trained neural networks show the proposed method is useful. In addition, the robustness about the parameters is discussed.

抄録全体を表示

In order to construct truly autonomous mobile robots, the concept of packaging is highly indispensable: all parts such as controllers, power systems, and batteries should be embedded inside a body. Therefore, implementing a controller on hardware is one of the most promising ways, since this contributes to low power consumption, miniaturization, and so on. Another crucial requirement in the field of autonomous mobile robots is robustness. That is, autonomous mobile robots have to cope with their unpredictably changing environments in real time. In this study, to meet these requirements the concept of Dynamically Rearrangeable Electrical Circuit(DREC) is proposed. In addition, we implement DREC onto FPGAs as physical electronic circuits by using the diffusion-reaction mechanism of neuromodulation which is widely observed in biological nervous systems. We developed the DREC for the peg-pushing task as a practical example. We confirmed that the physical DREC can successfully regulate the behavior according to the situation by changing its properties in real time.

抄録全体を表示

This paper is concerned with the development of a car-steering model for traffic simulation. Our focus in this paper is to propose a model of the steering behavior of a human driver for different driving scenarios. These scenarios are modeled in a unified framework using the idea of target position. The proposed approach deals with the driver’s approximation and decision-making mechanisms in tracking a target position by means of fuzzy set theory. The main novelty in this paper lies in the development of a learning algorithm that has the intention to imitate the driver’s self-learning from his driving experience and to mimic his maneuvers on the steering wheel, using linear networks as local approximators in the corresponding fuzzy areas. Results obtained from the simulation of an obstacle avoidance scenario show the capability of the model to carry out a human-like behavior with emphasis on learned skills.

抄録全体を表示

Abstract in this paper, a fast and reliable approach to the Traveling Salesman Problem (TSP) using the positively self-feedbacked Hopfield networks is proposed. The Hopfield networks with positive self-feedbacks and its collective computational properties are studied. It is proved theoretically and confirmed by simulating the randomly generated Hopfield network with positive self-feedbacks that the emergent collective properties of the original Hopfield network also are present in this network. The network is applied to the TSP and results of computer simulations are presented and used to illustrate the computation power of the networks. The simulation results show that the Hopfield networks with positive self-feedbacks has a rate of success higher than the original Hopfield network for solving the TSP, and converges faster to stable solution than the original Hopfield network does.

抄録全体を表示

Since the first neuron model was proposed, a lot of Neural Networks have been devised and been put into lots of practical uses. It is also true in the field of associative memory. Although so many useful memory models have been devised, there are still some problems, such as the limitation of storage capacity or too small attractor size to be stored.
In this paper, to solve the above problems, a novel associative memory is proposed. Its unique features are, (1) the memory network is obtained by training network parameters, (2) the size of the attractor of each stored memory can be controlled, and (3) some redundant nodes are introduced into the memory network in order to increase the storage capacity.
It is clarified from simulations that the proposed method can improve the memory functions, and can be applicable to the mutual associative memory easily.

抄録全体を表示

This paper proposes a multi-agent approach to power exchange planning for the electric power system. The proposed multi-agent system consists of Regional Network Area Agents (RNAGs), Branch Agents (BAGs) and Mobile Agents (MAGs). In this system, RNAGs are classified into three types of agents with respect to the net injection power to the regional network. The process is started when MAG is created from each demand network (DN), and MAG tries to find the required power from source networks (SNs) through the Internet. It is noted that the proposed multi-agent system can decide the power exchange planning using local information only. In order to demonstrate the capability of the proposed multi-agent system, it has been applied to a model network that consists of five RNAGs, seven BAGs, and two MAGs.
The simulation results show that the proposed system is effective and promising. The advantage of this approach lies in that no centralized control is required to guarantee the autonomous action, and appropriate power exchange planning is achieved by using decentralized approach.

抄録全体を表示

A linear precoding in time-varying multi-input multi-output(MIMO) channels is considered. We propose a channel extrapolation method for a linear precoding. A pecuriality of this method is to modify the pilot aided channel estimation for single-input single-output (SISO) systems and predict channel fading envelope using the polynomial extrapolation. The applicability of the proposed method is demonstrated by computer simulations.

抄録全体を表示

In this paper, we propose a new impulse noise filtering procedure from images using level detection. In our method, we use directional windows allowing continuity of edge. In our system, an window that is selected from multi-windows includes flat region. Thus, selected window has low variation in pixel values, and enhances the performance of restoration at edge. Extensive simulations indicate that proposed method performs significantly better than traditional methods.

抄録全体を表示

In an Ultrasonic cleaning tub, modes with high sound pressure, which are effective for cleaning, are called acoustic modes. In this paper, we show the effect of the presence of a washing object, when a stainless block which is fixed to the bottom of an Ultrasonic cleaning tub.

抄録全体を表示

In this paper, we present the performance analysis of two parallel Fast Fourier Transform algorithms, binary-exchange and transpose algorithms. These two algorithms were implemented and tested on a cluster of PCs using Message Passing Interface (MPI). The binary-exchange algorithm implementation was showing less processing and communication time than transpose algorithm.

抄録全体を表示

Wrong:(1) C. W. Ahn and R. S. Ramakrishna: “A Genetic Algorithm for Shortest Path Routing Problem and the Sizing of Populations”, IEEE Trans on Evolutionary Computation, Vol. 6, pp. 566-579 (2202)
Right:(1) C. W. Ahn and R. S. Ramakrishna: “A Genetic Algorithm for Shortest Path Routing Problem and the Sizing of Populations”, IEEE Trans on Evolutionary Computation, Vol. 6, pp. 566-579 (2002)

抄録全体を表示