Intrinsic modulation frequency responses of semiconductor laser diodes (LD's) are measured by using a novel frequency sweeper as the light source of the active-layer photomixing method. The modulation frequency responses of LD's up to 8GHz are measured avoiding the influences of the parasitic impedances. The frequency sweeper developed here employs only one distributed feedback LD, and is almost insensitive to ambient temperature fluctuations. Moreover, a high speed control of the beat frequency is possible, and the implementation of the lock-in detection is easily done with a suitable modulation pattern.
A two-way power divider which is composed of short-stubs and capacitances is presented. In a high frequency. this is a compact power divider because of a T network formed using series capacitances and shunt short-stubs. A design method is also shown for the case that the capacitance is accompanied with a little bit of lead lines. The experimental results of a trial power divider agree well with the theory. Moreover, in order to obtain a wide hand two-way power divider, this paper shows a design method of a power divider using a multi-sections of the T networks. A design example of two sections power divider exhibits a wide band characteristic such as a considerable improvement to the input VSWR is realized.
This paper proposes a method for applying Multi Transition Neural Network (MTNN) toward Fuzzy Multiobjective Coordination Combinatorial Optimization Problem (FMCOP). FMCOP has been puzzled over two kind of subjects which are pointed out until now. The one is how the membership function constructs, the other is how favorable solutions are obtained in reasonable time. To get satisfying membership function, interactive multiobjective programming which adjusts membership function interactively has been proposed. In examining whether the membership function is satisfying, plural favorable solutions are more desirable than the single optimal solution because comparison with solutions one another is made and this helps in judging membership function to be satisfying. Moreover, Combinatorial optimization needs considerable computation time, so fast approximation methods have to be developed to obtain solutions in reasonable time. MTNN has transition on Hamming distance greater than 2 by means of struggle for existence in the similar consideration of simulated evolution. It can effectively search for lower energy state, while escaping from local minima iteratively. Applying MTNN to FMCOP, the above two subjects can be resolved because it is capable of searching for plural favorable solutions in reasonable time. In consequence of numerical examples, MTNN has been verified to be effective for FMCOP.
We propose a dynamic thresholding method with feature-feedback in which feature information of the thresh-olded result is used to tune the thresholding in a proper state where a optimal thresholded result can be obtained. An original image is convolved with a two-dimensional Gaussian function first, and then the convolved image is used as a surface threshold at which the original image is thresholded. For the thresholded image, The selected feature value of extracted objects is calculated first and then, as the feedback information, is compared with the reference value. The space constant σ of the Gaussian is adjusted according to the error between the two values recursively until a proper result is given. The proposed method is applied to extracting nuclei in the glomeruli of human kidneys, where two features are selected as the feedback information; one is the average size of the extracted nuclei, and another is their size-number distribution. Experimental results are presented to successfully demonstrate the feasibility of the proposed method.
With rapid spread of EUC (End User Computing), RDB (Relational DataBase) construction and retrieval using GUI (Graphical User Interface) have become a major part of information network users. Since main activity in the RDB construction is its structure design, this should be carried out efficiently in order to shorten the construction term. Although the function of displaying entire structure (table structures and relations between tables) defined by users is provided as a “structure chart” display function in most RDBMS (Relational DataBase Management System) to support such activity, we can say that its quality is not high enough at present. For example, the chart contains many backward or intersecting arrows (relations between tables) and the tables placed are not well-balanced. Therefore, it is not easy to grasp the entire structure by the present chart. And this has been the cause of inefficiency in the RDB construction. In this paper, we propose a new method of making and displaying a high quality structure chart using BBM (Branch and Bound Method) and Fuzzy Inference as a solution to the above problems. Furthermore, we verify the effect of the proposed method in experiment.
It is the key step for face recognition systems to extract the facial parts from the complex backgrounds. In this paper, we propose a new method for the face extraction in the color complex backgrounds. By transforming color images from RGB color represention to YIQ color one, the orange-like parts including the face areas are enhanced in the original images, if the I-componet of YIQ color system is only used. The facial texture model based on the space gray level dependence (SGLD) matrices is applied to these images. Using this model, facial parts are detected as those regions which satisfy a set of inequalities. The weight coefficients in the inequalities are decided by the conventional method of learning. Using this textural model, we design a kind of scanning scheme for face detection in the complex backgrounds. The experiments show that this method could locate the face position in the complex backgrounds effectively.
In the vibration suppression control of the 2-mass system, the higher performance we demand, the higher the degree of the controller becomes. The controller is generally designed by a CAD system and is implemented with a microprocessor. But the microprocessor does not have enough precision to realize the result of the design by the CAD system. It is very important to establish implementation methods of the controller. Therefore, we show new implementation method of the controller based on the 2-DOF controller with a finite bit microprocessor. In this paper, first of all, we reviw the design algorithm based on the internal structure of the generalized 2-DOF controller. With this algorithm we can design the controller taking a robust stability condition into account. Secondly, we propose new coprime factorization of the 2-DOF controller in simple style. This factorization is based on the conventional PI controller, and thus can easily be applied to the conventional system. Thirdly, we show an implementation method of the controller taking account of the precision of the conventional microprocessor. This technique has been developed and used in the signal prossessing field. Finally, we show some experimental results with the microprocessor and verify the effectiveness of the proposed method.
In this paper, a method for design of a new product suited for the personal image is proposed. The method consists of two phases. In the first phase, a concept space is formed by using user's personal images of several existing products. In the second phase, a new product is designed from similar three products which are located near the position corresponding to required image in the concept space. This method can create products from the viewpoint of personal specification, because a concept space is formed based on each user's images of existing products. In this method, new product is designed from similar products by applying “operations.” “Operations” are defined as unit manipulations of some partial features of product. As a result, discrete parameters of design could be treated by this method. Using some simple pictures, the experiment to create new simple pictures suited for personal image is made. The concept space of the test subject is formed by his picture image, and some new pictures are created. Since new pictures produced in the experiment are suited for test subject's image, it is found out that this method could create new design suited for personal image.
There is a general method to simulate LC ladder filters, using equivalent inductances by GIC's. But, when an LPF is realized by this approach, the problems arise, such as DC bias instability and the influence of induced noise. First, a prototype LC ladder filter, which is obtained by connecting inductors in parallel with the input and output resistors of a conventional doubly terminated LC filter, is proposed, and the active simulation of this LC filter by GIC's, is presented. Second, a fifth-order elliptic LPF is designed. The measured transfer response shows a good agreement with the theoretical value. Finally, it is shown that the sensitivity of the proposed circuit is lower than that of the conventional circuits using GIC's.
It is aimed at in the present study to maintain a desired alveolar CO2 concentration by giving appropriate ventilation. It is, however, difficult to synthesize an adequate control system of respiration, because of a complicated respiratory regulation system which is dependent on its chronic change and difference of individuals. Hereby, effect of metabolic rate change is regarded as a characteristic change of dynamics resulting from parameter change of the respiratory regulation system whose input and output are ventilation and alveolar CO2 concentration, respectively. Based on the theoretical investigation of a simulation experiment, artificial respiration using fuzzy logic is proposed to maintain an alveolar CO2 concentration within an appropriate range of value. It is confirmed with its satisfactory stability and controlling performance by physiological experiments on healthy subjects.
In this paper a new subjective clustering method using fuzzy inference is proposed. Changing some parameters interactively, a user can reflect his/her knowledge or intuition for the clustering. The proposed method takes account of both (1) connectivity of data and (2) linearity of the data distribution. In addition, it represents shape of cluster by a membership function and uses fuzzy reasoning to reflect subjectivity of a user effectively. The proposed method is also effective not only for clustering but also for various purposes such as data analysis, assumption test, modeling, concept formation support systems, etc. The validity of the proposed method is confirmed by computer simulation.
PID control schemes have been widely used in most of process control systems represented by chemical processes for a long time. Indeed, they are easy for most of engineers in many industries to understand and for a various systems to use, whose properties are obscure. However, conventional PID control schemes have not been so available for the systems of which system parameters and the time delays are unknown. The purpose of this paper is to consider advanced PID control schemes in discrete-time systems with unknown time delays. Two control schemes, that is, compound PID control systems with a local feedback compensator and a pre-compensator, are proposed, and they are extended to explicit-type self-tuning PID control shemes. Furthermore, these methods are applied to a pressure control system in order to show the effectiveness of them.