An acousto-optic laser deflector was used as a shutter for high-speed imaging of laser interference fringes using an ordinary CCD camera. The exposure duration was set by the pulse width of the high frequency signal applied to the acousto-optic deflector. Changes in laser interference fringes due to an impulse discharge in air were obtained at an exposure duration of 4μs. By applying a sequence of high frequency signals with different frequency, the beam was deflected to four different angles at different times, allowing four interference images to be captured on a single video frame. This was used for time-resolved imaging of the interference fringe pattern.

View full abstract

A biothermal model of patient is proposed and verified for the brain hypothermia treatment, since the conventionally applied biothermal models are inappropriate for their unprecedented application. The model is constructed on the basis of the clinical practice of the pertinent therapy and characterized by the mathematical relation with variable ambient temperatures, in consideration of the clinical treatments such as the vital cardiopulmonary regulation. It has geometrically clear representation of multi-segmental core-shell structure, database of physiological and physical parameters with a systemic state equation setting the initial temperature of each compartment. Its step response gives the time constant about 3 hours in agreement with clinical knowledge. As for the essential property of the model, the dynamic temperature of its face-core compartment is realized, which corresponds to the tympanic membrane temperature measured under the practical anesthesia. From the various simulations consistent with the phenomena of clinical practice, it is concluded that the proposed model is appropriate for the theoretical analysis and clinical application to the brain hypothermia treatment.

View full abstract

The auto-correlation function graph figure considerably represents characteristics of image figures and outline shapes. This paper presents a method for classifying registered trademark images according to a function which has been made of the auto-correlation function. Most of registered trademark images are formed with a couple of outer figure or frame and inner figures or characters. Due to this, image figure patterns and the number of pixels have influence on the auto-correlation function for image patterns and its graph figure can be separated into three domains where each figure shows quite different characteristic. Considering the characteristic on each domain, the auto-correlation function can be exchanged into a set of three piecewise expressions: an exponential, a sign, and a third order polynomial expression. This transformed function is defined as the similar function in this paper. The point of this function is whether its graph figure is similar to the original auto-correlation graph figure. Experimental results with 1103 registered trademark images have shown that the proposed method has conducted the same classification as image classification by 20 people at a rate of 93.7%.

View full abstract

This paper presents a method for face recognition based on fuzzy clustering and parallel neural networks. Neural networks have been widely used for pattern recognition in various fields. However, the computing efficiency and the recognition rate decrease rapidly if the scale of neural network increases. In this paper, a new method of face recognition based on neuro-fuzzy system is proposed. In particular, the face pattern is divided into several small-scale parallel neural networks based on fuzzy clustering. Experimental results show the effectiveness of the proposed method.

View full abstract

A new automatic filter construction system for tracking an object is shown in this paper. Firstly, this system divides an input image with color features using mean shift clustering. And then, a new set of clusters are recombined into an optimal filter set which can extract inherent features. Furthermore, this system extract target using color features without the features of its shape and position. So even if the target object disappears in some frames, this system can detect it in the next some frames when the target appears again.

View full abstract

This paper proposes an efficient environment to develop plant control software (PCS) that controls the space plants for satellites or space stations. Generally, PCSs for space plants have been individually developed because each plant operates in different way and have strict restrictions on types, performances and admeasurements of the hardware to be used. Because of this individual development, the qualities of PCSs heavily depend on the developers’ abilities and experiences, and the development period tends to get longer. To solve this problem, we propose an efficient environment that make possible to develop high quality PCSs (1) by developing software components based on domain analyses of PCSs, and (2) by conveying design information accurately throughout the development processes. We have developed a prototype of this environment and applied to the actual developments of PCSs. As a result, we are able to convey precise design information throughout the development processes and have succeeded to develop PCSs that are fully conformable to our requirements. We are also able to predict the program size and the processing time of the PCSs within 10[%] errors at the initial phase of development. Moreover, we have developed most parts of the PCSs using software components. The reuse rate of software components is 65.5[%], and development time is cut down to 42.1[%] in comparison with the developments from scratch.

View full abstract

Recently, a general methodology for designing chaotic and hyperchaotic cryptosystems was developed using the control systems theory. Grassi proposed a nonlinear-observer-based decrypter for the state of the encrypter. If we can design the decrypter without the knowledge of the parameters of the encrypter, the chaos-based secure communication systems are not secure. In this paper, we design an observer-based chaotic communication system, which allows us to assign the relative degree and the zeros of its encrypter system. Moreover, under some conditions, we design an adaptive decrypter using the adaptive parameter adjustment law based on a Riccati equation in the case where the transfer function of the encrypter is of minimal-phase type.

View full abstract

In this paper, we propose an algorithm realizing minimum energy control for air conditioning. In the air conditioning system, control inputs are valve-open-value of cooler, heater and humidifier, and controlled variables are indoor temperature and humidity. That is, this system has three inputs and two outputs. The main step of the minimum energy algorithm is to switch two of the three actuators, depending on setpoints, for minimizing consumption energy. We apply this algorithm to a real plant, and show validity of the algorithm.

View full abstract

Neural networks are widely used for many fields including control problems. When we use a neural network as a controller, the direct inverse control scheme is the simplest approach. However, the scheme using the back-propagation method requires information, e.g. Jacobian, of an objective. If the objective is nonlinear, it will be difficult to know the information. Using the simultaneous perturbation learning rule, we can construct a neuro-controller, which is an inverse system of the objective or a proper controller, under the direct inverse control scheme without any information on the objective.
In this paper, we report a tracking control for a two link robot arm using a neuro-controller with the simultaneous perturbation learning rule. Details of the proposed control scheme are described. Some simulation results and a result for a real two link robot arm show that the control scheme is feasible and useful.

View full abstract

It is well known that the linear quadratic regulator of finite dimensional linear systems has the insensitivity property. In this paper, as a particular class of infinite dimensional linear systems, a neutral delay-differential system is considered as a plant. A class of linear quadratic regulators is constructed for the plant via a simple feedback law without real-time integral nor derivative operation. The feedback gain is calculated with a solution of a finite dimensional linear matrix inequality. First, it is shown that the regulator satisfies the circle condition. Then, its sensitivity against the parameter variation is evaluated. In the single-input case, it is done by calculating the absolute value of the sensitivity function. In the multi-input case, it is done by the method so called “comparison sensitivity”, where the sensitivities of the closed loop system and the open loop system are compared by some sensitivity indices.The procedure to investigate the property is a natural extension of one by Perkins and Cruz, where the finite dimensional case was treated. As a result, it is shown that the class of linear quadratic regulators of neutral systems has the insensitivity property as well as that of finite dimensional systems.

View full abstract

The purpose of this research is to make a calibration-free robot arm system which has a vision based controller. The system operation is carried out based on the simultaneous perturbation method.
Generally speaking, many robot systems involve a kind of vision-based controller which is able to control positions of a manipulator with explicit calibration of camera and robot parameters. However, these systems require much information about geometry of camera parameters and kinematics of the system. At the same time, calibration itself will be a time-consuming process as well. As a result, these systems become relatively complex. Moreover, if circumstances or camera parameters change, we have to recalibrate the system.
In this paper, we would like to propose a calibration-free robot control system using the simultaneous perturbation method. By using the simultaneous perturbation method, the system does not require any information on the robot or the camera. Actually, we can move the tip of the robot arm to goal point without any information on the system. Actual experiments show feasibility and usefulness of the proposed system.

View full abstract

Genetic algorithms are often well suited for optimization problems because of their parallel searching and evolutionary ability. Crossover and mutation are believed to be the main exploration operators. In this paper, we focus on how crossover and mutation work in binary-coded genetic algorithm and investigate their effects on bit’s frequency of population. According to the analysis of equilibrium of crossover, we can see the bit-based simulated crossover (BSC) is strong crossover method. Furthermore, to increase robustness of binary-coded genetic algorithm, multi-generation inheritance evolutionary strategy(MGIS) was proposed. Simulation results demonstrate the effectiveness of the proposed method.

View full abstract

So far, in associative memory search problems chaotic neural networks have constant synaptic weights to store patterns. In this paper, we propose a chaotic neural network(CNN) which has function typed synaptic weights to store patterns in order to make a better performance of the retrieval of the stored patterns. In stored patterns retrieval simulation, it is clarified that our proposed method is superior to the conventional method, that is, which has constant synaptic weights. Furthermore we propose an algorithm to calculate the mutual information in a CNN and show that the mutual information in the CNN, which are on the edge of chaos, gets the biggest values.

View full abstract

Nowadays there are several commercial electrical noses (ENs) applied in many applications, mainly in food and cosmetics industries. Most of them have been added with complicated mechanisms to control the measuring environment. Consequently, they are large in size and expensive. However, the reliability of those ENs can be achieved only at moderate levels. Therefore, a simple EN system with an effective method to analyze the data is proposed as an alternative way for classifying smells. The EN has not been added with a mechanism to control the measuring environment. Thus, the EN system is inexpensive, small and can be operated easily. However, a normalization method need to be utilized to reduce the effect of measuring environment. Then a method to select the representative training data for artificial neural networks (ANNs) based on a similarity index (SI) value is applied to reduce the training time. The results show the ability of the EN that is able to classify not only different kinds of smoke but also the same kind of smoke from different brands and different concentration levels quite precisely.

View full abstract

In this research, a pre-download method in the information service system using Bluetooth is proposed. The pre-download method is to determine the information to the visitor’s mobile in advance by using the movement prediction of the visitor. By using this method, the visitor can browse the information smoothly.
This method is evaluated by simulation in various room-layouts and walking routes. More than 92% of requested contents are pre-downloaded in case the memory size is large enough to download. The prototype of the information service system installed the proposed method has been developed. By using this system, the effect of the method is confirmed.

View full abstract

At present, the most commonly used satisficing method for multi-objective linear programming (MOLP) is the goal programming (GP) based method but this method does not always generate efficient solutions. Recently, an efficient GP-based method, which is called reference goal programming (RGP), has been proposed. However, it is limited to only a certain target point preference, which is too rigid. More flexible preferences such convex polyhedral preferences are preferred for many practical problems. In this research, a satisfactory effective linear coordination method for MOLP problems with convex polyhedral preference functions is proposed. The concept of the convex cone is used to formulate the convex polyhedral preference function and the existing lexicographic model of the reference point method (RPM) is integrated to ensure the efficiency of the solution of the problem. The formulated model can be solved by existing linear programming solvers and can find the satisfactory efficient solution. The convex polyhedral function enriches the existing preferences for efficient methods and increases the flexibility in designing preferences for decision makers.
In some situation, it is difficult for a decision maker to state a certain desirable level for each objective function. Applying fuzzy goal to capture the decision maker’s preferences has the advantage of allowing for vague aspirations, which can be considered as convex polyhedral preference functions. The satisfactory efficient linear coordination method can be applied to obtain an efficient solution, which is also close to the decision maker’s requirements.

View full abstract

In this paper, a VLSI chip of a discrete-time chaos circuit realizing a tent map is reported. The VLSI chip is fabricated in the chip fabrication program of VLSI Design and Education Center (VDEC). A simple structure enables us to realize the circuit with 10 MOSFET’s and 2 capacitors. Furthermore, the circuit which is designed by switched-current (SI) techniques can operate at 3V power supply. The experiment concerning the VLSI chip shows that the proposed circuit is integrable by a standard 1.2 μm CMOS technology.

View full abstract

We propose a method to recover the original sound waveform from the observed sound waveform without using the inverse filtering of the room transfer function. Reverberation is caused by the echo element of the original sound signal. In this study, short time frame signals subdivided reverberant speech by time windowing, and we attempted to reduce the echo influence by subtracting previous frame signals from the current frame signal. Weighing coefficient and adjusted time for the subtractions are determined by correlation between past and current signals. In the case that reverberation time is about 1 sec, the result of simulation shows good recovery of original speech.

View full abstract

In electric power facilities, a lot of analog meters without data transfer function are used. Field engineers have to read them manually in their daily work. In order to support the field engineers, this paper proposes an automatic reading technique of analog meters. Two features of this technology are automatic basing point compensation and automatic compensation of difference in position of the meters.

View full abstract