Organic semiconductor transistors are expected for flexible displays, infomation tags, etc. However, organic transistors have some disadvantages of current density and speed of operation, because of their high resistivity and low carrier mobility. To improve their performance, it is valuable to analyze the operation mechanism. We have compared characteristics of lateral- and vertical-type transistors using pentacene evaporated films. The pentacene lateral-type FET characteristics showed large variation due to fabrication parameters such as deposition rate, substrate temperature, and film thickness. It was found that evaporated pentacene thin films have large grain gaps, which disturb carrier flow along the channel formed pentacene film / SiO₂ interface direction and make FET characteristics instable. It is assumed that pentacene films are not suitable lateral-type transistors. Accordingly, we have fabricated vertical-type organic static induction transistors (SITs) using pentacene evaporated films. The pentacene SITs show a high speed operation (cut off frequency of 7.5 kHz) at relatively low voltages. Further optimization of SIT structure will improve the device performance.

View full abstract

Recently, much attention has been paid to novel photochemical vapor deposition using vacuum ultraviolet excimer lamps (VUV-CVD), because thin films can be deposited at room temperature. We show some characterization of films and discuss about the reaction mechanism using organic siloxane gases in VUV-CVD. VUV photons dissociate organic gases into radical reactants of Si-O and impurities such as C-H and O-H in the gas phase. The reactants finally condense on the surface into films included SiO₂ and impurities. The reaction is different by adding O₂. It is caused by the behaviors of ozone and activated oxygen that VUV photons dissociate O₂ and are generated. The deposition efficiency is improved both by the low substrate temperature and the higher light-intensity of an excimer lamp. The film deposition at room temperature in VUV-CVD that either thermal or plasma induced damages are completely avoided must become a novel technique in thin films preparations.

View full abstract

This paper describes how to develop a general-purpose LSI for analog signal processing for the signal obtained by variety of sensors. The LSI system includes eight independent elements and variety of signal processing circuit can be realized just by properly combining some of the elements functions. The eight elements are input-output full swing operational amplifiers, low noise operational amplifiers, monitor circuits, an absolute detector, filters, a current generate to adjust the filter cut-off frequency, an AGC, and reference voltage and current circuits. Among those, the input-output full swing operational amplifiers, the filters, and the absolute detector were newly designed by applying the special circuit techniques to make the circuits highly accurate. The conventional simple full swing operational amplifiers were mostly realized by C-MOS transistors, whereas the proposed circuits cannot help using the bi-polar transistors because variety of circuit elements are realized in the LSI. The LSI realized both by bi-polar and C-MOS transistors are more expensive than the LSI simply realized by bi-polar transistors. The gyrator techniques were newly employed to realize the filters by which the type of filters, i.e., HPF or LPF can be easily selected by the external connection of elements and the cut-off frequency can be easily adjusted by controlling the current to the circuit. The absolute detector does not include diodes from which the detector circuit is less-influenced by the temperatures and has the less-distortion characteristics.

View full abstract

This paper proposes a new type of multi-stratified comb-line filters of narrow pass band utilizing conductor-backed coplanar waveguides. It is shown through numerical simulations by means of FD-TD method that the filtering characteristics can be controlled effectively by introducing a grounding conductor with a window between the coupled two resonators. It is demonstrated that, by changing the dimensions of the window on the grounding conductor, we can change, without affecting the odd mode, the input susceptance of the even mode independently and that we can adjust the frequency of attenuation pole. An interesting feature of this filter is that an additional window on the grounding conductor makes another attenuation pole appear around the passband.

View full abstract

Ultra-short optical pulses synchronized to an electrical clock are required in ultra-high bit rate soliton transmission and optical signal processing systems. Use of sinusoidally-driven electroabsorption (EA) modulators is described to generate transform-limited optical pulses. We model pulse generation in sinusoidally-driven lumped-electorode EA modulators, defining a new device figure of merit. Generated optical pulse width is calculated as a function of extinction ratio slope for applied voltages 3dB bandwidth and sample length, and calculated results are compared with those reported for various modulator structures that include bulk modulators, InGaAs/InAlAs, and InGaAsP/InGaAsP multiple quantum well (MQW) modulators. This model is useful for designing and using EA modulators in optical short pulse applications.

View full abstract

MOCVD method is generally adopted as the epitaxial growth method of AlGaInP red-light emitting laser diodes. Authors have succeeded in developing high-power red-light emitting laser diodes having long life by solid-source MBE method for the first time in the world. Contents in this paper are the explanation of the MBE system used in this study, the growth condition of high quality crystal, and fabrication process anDHigh power characteristics of loss-guide lasers and real-guide lasers. In the wavelength of 656nm, long life more than 2500 hours is obtained under the pulse condition of 70 mW, 70°C. This laser can be used as a light source for DVD-R/RW and DVD-RAM.

View full abstract

In this study, we examined consumption of processing resources in face recognition on working memory. Selective interference is occurred by dual-task method with matching-to-sample. We assess processing delay and correct percentage of task that time. As recognition categories, we used simple figures, languages, objects, scenes, and faces (considering everyday life and vascular dementia). By experimental result, we understood consumption of processing resources in face recognition is the largest among other categories on working memory. Using the relation of processing resources consumption between face recognition and other recognitions on working memory, we hope that assessing of vascular dementia noticed frontal lobe dysfunction.

View full abstract

We analyzed the feedback relationship between short-term fluctuations in heart rate and blood pressure in healthy persons and heart failure patients. Parameters derived from the feedback relationship between heart rate and blood pressure have been proposed. The purpose of the present study is to apply these parameters in estimating autonomic function or measuring physiological and mental workload.
Electrocardiographs and beat-to-beat blood pressure were recorded in supine position at rest. The blood pressure was measured using arterial tonometry. The R—R interval and systolic blood pressure were fitted to two-dimensional autoregressive models, the relative power contribution in the frequency domain was calculated. The proposed parameters are the power contribution in the low-frequency range ( 0-0.15 Hz ) [ RS_LF, SR_LF ] and the power contribution in the high-frequency range ( 0.15-0.5 Hz ) [ RS_HF, SR_HF ].
RS_LF was significantly different between healthy persons and heart failure patients ( p<0.01 ). This parameter can be used to estimate autonomic depression caused by aging and heart failure.
There were no correlations between the proposed parameters and the usual indices to evaluate autonomic function. It is considered that the proposed parameters can be used to evaluate physiological states that cannot be evaluated using existing methods.

View full abstract

In case of flood, earth and sand at the bottom of bridge piers is sometimes washed out and thus yields unstableness of the bridge piers. Therefore, monitoring of bridge piers is of great importance in keeping the safety operation of trains. The paper proposes a simplified monitoring system which analyzes the vibration of the bridge piers generated by an impact with a light hammer. The system uses an acceleration pickup set on the top of the bridge piers to extract the vibration signal and makes use of system theoretical methods such as Kalman filter and maximum likelihood method.

View full abstract

The authors focused on the applicability of capacitance sensors as a human body detecting/monitoring means to realize high-performance equipment to assist in caring for the aged. As the human body contains a high percentage of water, it has a high specific inductive capacity to be detected, that sets it apart from objects such as plastics. This report describes the positive result obtained by using FEA (Finite Element Analysis) to determine how capacitance is affected by parameters such as the interval between measurement electrodes, the distance from an object, and differences in specific inductive capacity of objects with respect to a pair of square electrodes arranged on a plane. Capacitance was calculated by preparing a 2-dimension model of plane type capacitance sensor utilizing the general-purpose FEA software ANSYS. As a result, it was found that increasing the interval between measurement objects enables them to be detected from a greater distance. This distance is approximately equal to that between electrode centers, and is an important factor for detection limit distance defining the performance of plane type capacitance sensors. Furthermore, this paper shows that the sensor is impervious to intervening material, which may exist between the sensor and object, and that the sensor is capable of detecting the presence of a person, due to the high dielectric constant of the human body.

View full abstract

This paper proposes a distributed algorithm for adjusting the signal polarity of the control networks for building air-conditioners. As field connection work of the two wires to the controller may be straight or cross, each controller has both plus and minus polarity circuits. It starts the algorithm in the arbitary sequence, and then negociates each other to select the plus/minus polarity circuits in order to adjust the signal polarity. A Time Petri Net model was constructed for verification of the proposed algorithm. The model describes the required specification of the system, straight/cross connection of the controller to the network bus line and the polarity adjustment algorithm. The efficient Time Petri Net verification tool made it possible to verify the model.

View full abstract

Authors have been studying Fluency Information Theory that generalizes Shannon’s sampling theorem and its applications. Among the practical application of the research, the Fluency DAC that is developed as the Digital-to-analog converter for CD audio could have received objective valuation including receipt Golden Sound Award in 1988. In recent, DVD-Audio that deal with maximum sampling rate of 192 kHz has appeared. Due to the introduction of DVD audio that requires four times the sampling rate of nowadays CD audio, the request for developing a new Fluency DAC for DVD audio was initiated. From such requirements, the research for developing the Fluency DAC for DVD-Audio has been started. The result of the research could revive awards in local contest in Japan audio apparatus at 2000 and 2001. As the initial report on our project in developing the Fluency DAC that is capable of dealing with a maximum sampling rate of 192kHz, in this paper we aimed to derive the sampling function that acts as the impulse response for such a D/A converter.

View full abstract

In this paper, the authors propose a method for detecting eyes from the facial image by using the genetic algorithms. This method is not influenced against the image obtainment environment, the expression of the face and the change in the directions of the face. The eyes are detected by using the genetic algorithms; the outline of the eye is considered two ovals and the pupil of the eye is considered the circle. Directions of eyes are detected from the relations of the outline of the eye and the eye at the end.

View full abstract

This paper presents a method to use wavelet analysis for speech coding and synthesis by rule. It is a coding system where LSP residual signal is transformed into wavelet coefficients. As wavelet analysis is implemented effectively by filter banks, our method is featured to require less computation than multipulse coding and others where complicated prediction procedures are essential. To achieve good quality speech at low bit rates, we verified to allocate the different bits onto the wavelet coefficients, with more bits in lower frequencies, and less in higher. The synthesized speech by Haar wavelet with 16.538kbits/s has nearly same perceptual quality with 6 bits μlog PCM (66.15kbits/s). We are convinced that coding method of LSP residual signals using wavelet analysis is an effective approach to synthesize speech.

View full abstract

This paper proposes a new on-line parameter estimation method with radial basis function networks for time-varying linear discrete-time systems. The time-varying parameters of the system are expressed with the radial basis function networks. These parameters are estimated by the nonlinear optimization technique, and the setting rules of the initial values in the optimization are proposed. The system parameters are usually unknown because they are changed by the circumstance conditions. Then, it is reasonable that the structures of the radial basis function networks are regulated according to the change of parameters. The minimum description length criterion studied in the encoding theory is applied to select the network structures. It is demonstrated in digital simulation that the proposed on-line estimation method succeeded to reduce the computaion time extremely, for time-varying parameters system.

View full abstract

PID control schemes have been widely used for various process control systems for a long time. However, since such processes have relatively large time-delays and uncertainties caused by modeling errors and system fluctuation, it is non-trivial to choose a suitable set of PID parameters. In this paper, a design scheme of self-tuning PID controllers is proposed for unknown processes with time-delays. This scheme is equipped with a parallel feedforward compensator to compensate for the time-delays. The PID control system is then designed for the augmented system. The PID control parameters are calculated adaptively based on on-line process identification and the relationship between the minimum variance control law and PID control law.

View full abstract

In this paper we present the algorithm for the algebraic Riccati equation with indefinite sign quadratic term for the singularly perturbed system by means of the Newton’s method. The proposed algorithm is very efficient from the numerical point of view because the algorithm is the quadratic convergence and the algebraic Riccati equation is solvable even if quadratic term is an indefinite sign. The main contributions of this paper are as follows. First, the sufficient condition corresponding to the perturbation parameter such that the proposed algorithm is quadratic convergence is presented. Second, the method which is based on the linear equations method for solving the algebraic Lyapunov equation of singularly perturbed system is given. In order to show the effectiveness of the proposed algorithms, a numerical example is included.

View full abstract

Motivated by a general Hurwitz matrix expression involving a triplet of matrices, its Schur counterparts are first derived. This time, they are expressed by a pair of matrices satisfying a certain nonlinear or norm condition. It is then shown that the results can find applications in Schur stability analysis of a polytope of matrices. Using the obtained expessions, two kinds of quadratic Lyapunov functions are proved to work for the polytope:a fixed quadratic function and a parameter-dependent quadratic function. The first kind gives a well known extreme point result on quadratic stability of polytopes of matrices, while the second yields a new result for the stability test of the polytope.

View full abstract

In this paper, we propose a design method of neurocontrollers (NCs) evolved by a genetic algorithm (GA) for the backward movement control of a truck system with two trailers. The difficulty and complexity of the backward movement control depend on the number of connected trailers and the angular limitations of the trailer-truck system. In order to search for the best neurocontroller more quickly and effectively, for the two-trailer system which has extremely small angular limitations, we propose a modified GA which adaptively changes the number of offspring and the mutation rate according to the diversity of NC population. The simulation results show that the modified GA significantly improves the search performance. We apply the control method not only to computer simulations but also to experiments of a small-scale real mechanism. The results of both show that the control method is highly effective.

View full abstract

Recently, the extended correlation least mean squares(ECLMS) algorithm have been proposed to solve the double-talk problem in the echo canceling system. The characteristic of ECLMS algorithm is to utilize the correlation functions of the input signal instead of the input signal itself. ECLMS algorithm, based on the linear filter can separete noise signal from observed signal. This algorithm can not be used for nonlinear time series. To solve this problem, we have proposed the ECLMS algorithm for second-order Volterra filters. However, The number of operations required to consider all combination increases exponentially with the highest order term of the Volterra filter.
So, we propose Neural networks(NN) system for nonlinear time series estimation using ECLMS algorithm. It is easy to extend for the highest order term of the Volterra filter. Some numerical examples are presented to illustrate that the proposed method can work well for estimating nonlinear time series with ARCH errors.

View full abstract

Backpropagation, one of the most popular learning algorithms in multi-layered feedforward neural networks, suffers from the drawback of slow convergence. Several modifications have been proposed to accelerate the learning process using different techniques. In this paper, a new cost function expressed as exponential of sum-squared or Log-likelihood is proposed. Weight update using this modification varies the learning rate parameter dynamically during training as opposed to constant learning rate parameter used in standard Backpropagation. Simulation results with different problems demonstrate significant improvement in the learning speed of Backpropagation algorithm.

View full abstract

In this paper we address a new problem in reinforcement learning. Here we consider an agent that faces multiple learning tasks within its lifetime. The agent’s objective is to maximize its total reward in the lifetime as well as a conventional return in each task. To realize this, it has to be endowed an important ability to keep its past learning experiences and utilize them for improving future learning performance. This time we try to phrase this problem formally. The central idea is to introduce an environmental class, BV-MDPs that is defined with the distribution of MDPs. As an approach to exploiting past learning experiences, we focus on statistical information (mean and deviation) about the agent’s value tables. The mean can be used as initial values of the table when a new task is presented. The deviation can be viewed as measuring reliability of the mean, and we utilize it in calculating priority of simulated backups. We conduct experiments in computer simulation to evaluate the effectiveness.

View full abstract

The saddle type of fixed points are embedded in the chaotic trajectory from complex systems, and the estimation of these points in unknown nonlinear systems is significant. At first, we present a coordinative method to construct a nonlinear model approximating time series from unknown nonlinear system and to estimate its fixed point by using a neural networks and nonlinear optimization techniques.
Finally, we apply this presented method to OGY method which was presented as one of the chaos control. We can stabilize chaos to the fixed point by using this method. However, in order to construct feedback rules in OGY method, it is necessary to obtain the system equation and its fixed point explicitly. To overcome this problem, we apply the proposed coordinative method. As a result of this application, we can stabilize chaos to its fixed point by using only the time series from the nonlinear system.
We also apply the method to the global optimization method by chaos series. In our method, the global optimum can be estimated by using the chaos dynamics without stabilization by the chaos annealing.

View full abstract

Automatic categorization is the only viable method to deal with the scaling problem of the World Wide Web (WWW). In this paper, we propose a news web page classification method (WPCM). The WPCM uses a neural network with inputs obtained by both the principal components and class profile-based features (CPBF). Each news web page is represented by the term-weighting scheme. As the number of unique words in the collection set is big, the principal component analysis (PCA) has been used to select the most relevant features for the classification. Then the final output of the PCA is combined with the feature vectors from the class-profile which contains the most regular words in each class before feeding them to the neural networks. We have manually selected the most regular words that exist in each class and weighted them using an entropy weighting scheme. The fixed number of regular words from each class will be used as a feature vectors together with the reduced principal components from the PCA. These feature vectors are then used as the input to the neural networks for classification. The experimental evaluation demonstrates that the WPCM method provides acceptable classification accuracy with the sports news datasets.

View full abstract

We have fabricated field-effect transistors using TMPD-C_nTCNQ LB films and attempted to control the conductivity of the charge-transfer (CT) complex layers by gate voltages. Dependence of FET characteristics on acceptor molecules (C_nTCNQ; n=15,18), the stacking structure of the LB films and the number of layers was investigated. The FET characteristics strongly depended on number of CT complex layers. The difference in the form of LB films seems to originate from the strength of molecular interaction depending on the alkyl chain length.

View full abstract

In this paper, a new algorithm is presented for extraction of furrows from facial image. In order to extract furrows effectivery, the threshold values must be set place by place automatically. From this point of view, automatic setting algorithm of the threshold values is presented in this paper.

View full abstract

Over the past decade it has been widely recognized that not only the controllers implemented but also their embodiment, such as robots’ body, significantly influence the emergence of intelligence. In spite of its importance, still few studies have explicitly investigated this mutual interaction. In light of these facts, in this article, the interaction between controllers and their embodiments is investigated. To do so, a synthetic approach is employed. More specifically, construction of a locomotion controller that enables a passive dynamic walking (PDW) robot to walk on a flat terrain is taken as a practical example. In this study, so-called staged evolution is conducted for this purpose. First, body parameters (e.g. mass, length) of a biped robot are evolved in order to elicit PDW. After obtaining acceptable embodiments, their CPG (Central Pattern Generator)-based controllers are evolved. Simulations have been carried out and the results have shown that “appropriate” embodiments allow us to significantly increase the efficiency of developing their controllers.

View full abstract

This study proposes a neural network model which can recognize contact characters by only learning the standard characters. This paper describes how to experimentally recognize contact character on standard character and recognize the standard characters from the contact character.

View full abstract