Recently, the research and development of the system which is able to obtain the relaxation effect has been actively performed. Most of the moderns have some stresses and require the request for healing, and the early development of the relaxation system is expected. For example, the system which sounds and images are presented to the user, the relaxation effect is judged from the vital signals of heart beat, pulse rate and so on, and the contents of presentation is tuned to raise its effect has been developed. In this research, we aim to construct the automatic tuning system to obtain the relaxation effect by presenting the sound to the user and judging the relaxation degree based on the brain wave information. In this paper, as an advance research on the construction of this system, we try to analyze the features of the brain wave measured in a different state and to find the index that judges the relaxation degree.

View full abstract

In recent years, animal-assisted therapy (AAT), which makes use of the healing effects of animals has attracted attention. However, it is difficult to employ AAT in such facilities because of the risks of the spread of infection from animals to patients and the necessity of proper animal training. Thus, robot-assisted therapy (RAT), in which robots resembling animals are used instead of real animals, is important for patient safety. In RAT, it is important that the robot imitates the motions of a living animal, especially essential motions, such as walking, running, and so on. This paper presents a method for the generation of animal gait in quadrupedal robots. In this study, we employed AIBO as an experimental quadrupedal robot and generated the gait of the robot on the basis of an animal's gait. We generated the quadrupedal gait of AIBO using both the optimum orbit of the mono-leg and an animal's gait, classified as the gait of a walking dog based on zoology. Furthermore, minor deviation of parameters for each joint using chaotic time series analysis was corrected to realize the stable gait on the ground.

View full abstract

In this research, we propose a cooperative behavior control method which realizes the communication between robots based flexible decision enabled by using fuzzy inference. Using this method we aim to construct the system with the efficient task distribution ability in large complex environments. In order to confirm the effectiveness of this method, we performed several soccer robot simulation experiments in different conditions.

View full abstract

This paper proposes a method of human states estimation based on sensor networks. From a point of view of human state estimation, most of the previous researches depend on given knowledge, such as template matching. However, it is difficult to adjust flexibly to complicated environments such as houses or offices. Therefore, temporal features extractions and learning methods are very important. We propose a learning method of fuzzy spiking neural network based on a time series of measured data. Furthermore, we discuss the effectiveness of the proposed method through an experimental result in a living room.

View full abstract

In this paper, we propose a self-sufficiency model using a multinomial logit model (MLM) to control actions of agents. We thought that the agents can choose the flexible and appropriate actions, because a multinomial logit model is a kind of mathematical models, and suitable to model human's "judgment."

View full abstract

Reinforcement learning is a technique developed for a single agent. When we use it for the cooperative behavior in multi-agent environment as it is, the problem like how much reward should be distributed to each agent in order to learn correctly will be generated. Therefore, in this research we propose a reward distribution method called multi-agent profit sharing (MAPS) to solve the reward distribution problem occurred in multi-agent reinforcement learning. In this method, we evaluate all the behaviors of each agent by an individual behavior evaluation and a cooperative behavior evaluation using fuzzy rules. We compute the separate individual and cooperative contribution degree based on the behavior evaluation and distribute the reward according to the contribution degree. Using the cooperative contribution degree we evaluate all the behaviors from the overall system and construct the multi-agent system which agents are able to learn the cooperative behavior and group strategy efficiently.

View full abstract

This paper proposes a competitive facility location between two decision-makers (DMs), one of who has a priority to decision-making on a network with demands whose sites are uncertainly and vaguely. For representing such uncertainness and vagueness, the distance between a demand's site and one point in the network is given as a fuzzy random variable. Then, an optimal location problem for the DM with the above priority can be formulated as a Stackelberg location problem (SLP) with the fuzzy random variables. By using the alpha-level set, expectation, and variance, the SLP is reformulated as a version of conventional Stackelberg location problem on a network. Although previous study has shown theorems for the optimality of the SLP on tree network, this paper shows theorems for the optimality of the SLP on general network.

View full abstract

A simulation study is executed with respect to an end-of-period discount for daily perishable products. In case that supplied products will not be sold out by end-of-period, the sales floor manager sometimes sells the products in a discount price in order to increase the revenue of the period. The reference price of consumers for the products is consequently declined and some consumers would not purchase the products at a regular sales price. It is important for the manager to take the reference price effect into account so as to improve long-term profit. This paper formulates the end-of-period discount problem within a framework of dynamic programming. Optimal pricings are derived in a simulation study to estimate the influence of inventory distribution on the optimal pricings.

View full abstract

First of all, we discuss the generalization of quasi-concavity for membership functions defined on n-dimensional Euclidean space in this paper. Generally, quasi-concave functions are defined by using the minimum operation. Whereas we give the concept of the more generalized quasi-concave function by adopting some conjunctive aggregation function instead of by using the minimum operator. Then we derive some properties of the more generalized quasi-concave functions which we proposed. Next, we focus the plural membership functions defined on n-dimensional Euclidean space. Besides, we define a fuzzy mathematical programming problem whose objective function is an aggregation function on the range space of the plural membership functions. An optimal solution of the fuzzy mathematical programming problem is called a compromise solution of the problem by Ramik and Vlach. Finally, we derive the properties of the compromise solutions by means of features of the more generalized quasi-concave functions.

View full abstract

In this paper, we focus on hierarchical multiobjective linear programming problems with random variable coefficients where multiple decision makers in a hierarchical organization have their own multiple objective linear functions together with common linear constraints. In order to deal with objective linear functions and linear constraints with random variable coefficients, p-Pareto optimality concept is introduced. After each decision maker specifies his/her own decision power and reference probabilistic values, the corresponding candidate of the satisfactory solution is obtained among from p-Pareto optimal solution set on the basis of linear programming. Each decision maker updates his/her own decision power and/or reference probabilistic values according to the two rules, until the satisfactory solution is obtained. An interactive processes are demonstrated by means of an illustrative numerical example.

View full abstract

This paper considers a portfolio selection problem with Corporate Social Responsibility (CSR). In order to present CSR as a numerical value, the fuzzy theory is introduced. Using the mean-absolute deviation model and fuzzy goals for multi-objective functions, the proposed model is transformed into a linear programming problem. Furthermore, providing some numerical examples in the current financial market, the evaluation of the proposed model is performed.

View full abstract

We decide and execute our action from many types of environmental information in our daily lives even if we are not conscious of being guided. The human action induced from the opposing person's movement is the collision avoidance of passing each other. In collision avoidance, we chiefly judge the avoidance direction from visual information. Especially, it is important to get the information from oncoming person's body part and avoidance timing in each other. Then, we make an experiment to judge the avoidance direction by watching the masking movie of oncoming person's body part. By evaluating this judgment time, it was clarified oncoming person's body part is leg in collision avoidance. Next, it especially paid attention to oncoming person's leg, and the relation between the walking cycle and leg position in avoidance judgment is evaluated. From this result, the avoidance judgment is possible because the traveling direction can be controlled by the leg when the leg is lifting and landing. It was clarified that oncoming person's walking cycle is important in the action decision in collision avoidance. So we propose the action decision model based at the walking cycle from these results.

View full abstract

The authors recorded nineteen-channel event-related potentials (ERPs) during recognition of one type of Japanese character; Hiragana (one type of phonetic characters). By field-sequential stereoscopic 3D display with liquid crystal shutter, a word and a non-word were simultaneously and independently presented to the left (right) and the right (left) eyes, respectively. Each word consists of three Hiragana characters. Three subjects were instructed to press a button after they understood the meaning of the visual stimuli after 3000 ms poststimulus. Equivalent current dipole source localization (ECDL) with three unconstrained ECD was applied to the ERPs. The ECDs were localized to the Wernicke's area at around 600ms. The latter ECD for one left-handed subject was localized at the Wernicke's homologue. After that ECDs were localized to the prefrontal area, the superior frontal gyrus and the middle frontal gyrus. Then at around 800ms, the ECDs were localized to the Broca's area and after that ECDs were localized again to the Wernicke's area and then to Broca's area.

View full abstract

This paper presents a boosting-based approach to multi-class classification for brain-machine interface. First, to classify multi kinds of image, we discuss a way to acquire a number of local frequency features from electroencephalogram (EEG) signals. Next, a new boosting-based approach to multi-class EEG classification is developed by utilizing local support vector machines according to the local frequency features. The utility of the proposed approach will be directly presented at the symposium.

View full abstract

Brain-computer interface(BCI) and brain-machine interface(BMI) have been come into the research limelight. The outer computer and machine are controlled by brain activity and the discriminated boundary, which are detected with near-infrared spectroscopy(NIRS) or electroencephalograph(EEG). In this paper, we propose a new boosting algorithm for BCI using probabilistic data interpolation. In our model, interpolated data are generated by probabilistic distribution and assorted around errors instead of weights in the conventional Adaboost. By the interpolated data, the discriminated boundary is identified to control the outer machine effectively.

View full abstract

Recently, BCI(Brain-computer interface) and BMI(Brain-machine interface) come into the research limelight. However, nonsynchronous spontaneous action potentials and evoked action potentials exist contain brain signal, and we need an interface model between brain and machine for control and stability. We have already proposed collaborative learning system consisting of reinforcement learning and brain signal. Brain signal is interpreted as a deliberate assignment of the subject, and we utilize reinforcement learning in control and stability for BCI. In this paper, we discuss the usefulness of collaborative learning for BCI using reinforcement learning. We first design the collaborative learning system with near-infrared spectroscopy (NIRS), and apply it to maze problem. In addition, we discuss the comprehensive evaluation of collaborative learning system in terms of difficulty of the problem, precision of the problem and the mental load to the subject, and show the usefulness of the proposed system.

View full abstract

The P300 speller is one of the BCI applications, which allows users to select letters just by thoughts. However, due to the low signal-to-noise ratio of the P300, signal averaging is often performed, which improves the spelling accuracy but degrades the spelling speed. The authors have proposed reliability-based automatic repeat request (RB-ARQ) to ease this problem. RB-ARQ could be enhanced when it is combined with the error correction based on the error-related potentials (ErrPs) that occur on erroneous feedbacks. Thus, this study aims to reveal the characteristics of the ErrPs in the P300 speller paradigm, and to combine RB-ARQ with the ErrP-based error correction to improve the performance further. The results show that the performance of the P300 speller could be improved by 40 % on average.

View full abstract

This paper discusses a brain-machine interface (BMI) to achieve the so-called training-free concept and real-time classification. To accomplish the purpose, we construct auto-regressive models with exogenous variables (ARX models) for EEG dynamics and propose a way of classifying two kinds of image from prediction correlation of two ARX models. The classification results show the possibility of a training-free BMI construction and real-time classification. Finally, we apply our BMI system to control of an electric wheelchair and its experimental result demonstrates the utility of our developed BMI.

View full abstract

This paper presents the learning based self-organized additive fuzzy clustering method and its application of the use of electroencephalogram data. The proposed method combines the learning process to the conventional self-organized additive fuzzy clustering method by using inner product between a pair of degree of belongingness of objects. By learning the status of the noise in each process of iteration of the algorithm, the proposed method can obtain a more adaptable result. In future work, we believe that this technique can be useful for developing a man-machine interface in which the obtained classification result by the proposed method is used for the discrimination process of human's thoughts.

View full abstract

The Brain-Computer Interfaces (BCIs) control a computer or machine based on the information of human's brain, and the P300 speller is one of the BCI communication tools. The P300 speller discriminates a character after averaging the plural EEG data to improve the accuracy. Whereas reliability-based automatic repeat request (RB-ARQ) is an error control method designed for BCIs, which makes a user keep thinking until a given reliability is satisfied and can improve the accuracy of BCI with a small loss of the discrimination speed. And, this paper proposes RB-Selective Repeat ARQ (RB-SR-ARQ) which selectively requests a user to re-send the data based on the reliability of each data. This paper applies RB-SR-ARQ to the P300 speller, and then it shows that the time required for thought discrimination is able to be reduced while same accuracy is preserved.

View full abstract

In this paper, new classification method of situations in sensing environment using web camera device for cooking learning support was proposed. For usability and portability of learning environment, in this system valuable information, such as user's hand position, state of cooking devices and position of ingredients, are measured from marker's move. Moreover, by positions of marker are classified into divided area (a part of kitchen), it is expected that learning situations can be easily compared with collect data by learning support system. In addition, user can be indicated virtual environment that integrated movie and 3D objects (such as cooking devices and ingredients) by using Head Mounted Display. This integrated visualization support can help user learn cooking process with realistic sensation. This experimental result indicated that proposed method can measure the position of markers in working environment with high accuracy. Furthermore, it is shown that the series of area calculated based on user's cooking process can be accurately estimated.

View full abstract

A lot of study support that uses e-learning including the teaching material distribution for the lecture and self-study is done. The quiz is one of the teaching materials for self-study. However, it takes time to make the question. And make notes is one of the study method. Notes is learning material at earlier study, and design-based learning environment learned based on the note is built. In the present study, to make questions automatically based on the teaching material that the learner designed, it proposes the approach for operating the data base of an existing e-learning system directly.

View full abstract

The purpose of this paper is to construct a lecture note generating system with free layout using a laptop PC. The students can put contents on the interface of the proposed system freely. The information of inputted contents is stored in the database. The texts in the contents are morphologically analyzed to obtain the important keywords. The supplemental explanation that is searched on the Web for the keyword is also put on the interface. Both the inputted contents and the obtained information on the Web are reconstructed by extracting a part of them and arranging together. The system is constructed as the Web application, which enabling us to share the note in a community. By realizing these functions, we construct the system without losing advantage of a notebook and with convenience of PC.

View full abstract

The infrared optical source was shed on the fingertip and the intensity of the transmitted light or the reflection light was measured. The volumetric change of the bloodstream was estimated from these measurements. The acceleration plethysmogram wave data were obtained by the 2nd differentiation of the original measured waves. The results of the waves were classified by the Self-Organizing Maps (SOM) by 2-dimensional (plane SOM) and 3-dimensional (spherical SOM). The display of the clinical examples was also successful related to the Acceleration Plethysmogram Analysis.

View full abstract

This paper shows the performance for diagnosis of fundus images using the subspace method. Feature extraction was conducted with three kinds of image (R, G and B pixel), and the feature vector and the subspace dimension for recognition were determined. Afterward, a series of analyses on the accuracy were conducted. The recognition accuracy of the fundus image by using the subspace method was compared with the results obtained by two methods: Learning Vector Quantization and Multi-Layer Perceptron. In the experiments, consequently, a maximum accuracy rate of 75.2% was obtained by using the subspace method, in which the accuracy was the highest performance among three methods' accuracy.

View full abstract

In this paper, a diagnosis support system for mucous membrane diseases in oral cavity, in which a shape of vitiligo is judged as lacy or not-lacy from an intraoral image, is proposed. To realize good discrimination, it is important to extract vitiligo area with high accuracy. To segment vitiligo from other area, a probabilistic relaxation with teacher signal is employed. From some simulations, an effectiveness of the proposed method is verified.

View full abstract

This paper introduces integrated software of computer-aided diagnosis of acute coronary syndromes (ACS) by using an intravascular ultrasound (IVUS) image. The software contains some powerful functions for the diagnosis of ACS such as fully automatic plaque boundary extraction function in the intravascular ultrasound (IVUS) image by using anisotropic-diffusion-based pre-processing and Takagi-Sugeno (T-S) type fuzzy inference with a weighted separability measure, fully automatic tissue characterization function by using multiple k-nearest neighbor classifier, 3-dimensional visualization function of the characterized coronary plaque, and so on. The effectiveness of the produced software is verified by the experiments using the real IVUS images.

View full abstract

Since the single input rule modules connected fuzzy inference model (SIRMs model) is proposed by Yubazaki et al., many researches on the extension of the SIRMs model have been reported. This paper shows the equivalence between the fuzzy funcitonal SIRMs inference model and fuzzy functional inference model, in order to clarify the property of generalized SIRMs inference models.

View full abstract

This paper addresses equivalence of type-2 fuzzy inference model. It first presents fuzzy inference models: the type-2 fuzzy inference model and fuzzy functional inference model. Next, two fuzzy inference models are shown to be equivalent.

View full abstract

In data mining approach, the quantitative attributes should be appropriately dealt with as well as the Boolean attributes. This paper presents a fast algorithm for extracting fuzzy association rules from database. The objective of the algorithm is to improve the computational time of mining for the actual application. In this paper, we propose a basic algorithm based on the Apriori algorithm for rule extraction utilizing redundancy of the extracted rules. The performance of the algorithm is evaluated through numerical experiments using benchmark data. From the results, the method is found to be effective in terms of computational time and redundant rule pruning.

View full abstract

In order to realize intelligent agent such as autonomous mobile robots, Reinforcement Learning is one of the necessary techniques in the control system. It is desirable in terms of knowledge or skill acquisition of agents that reinforcement learning is based only upon rewards concept instead of teaching signal. However, there exist many problems to apply reinforcement learning to actual problems. The most severe problem is huge iterations in learning process. Our motivation is to utilize appropriately instructions that we can give to the reinforcement learning agent along with main rewards in order to haste the learning process and to attain valid learning performance for preparation of segmentation. In this study, we propose an instruction approach for reinforcement learning agents based on sub-rewards and forgetting mechanisms. Through numerical experiments of the grid world task and the mountain car task, we show validness of the proposed approach in terms of learning speed and accuracy.

View full abstract

A management and economic system have complex systems. For this reason, it is hard to handle a management and economic area's data. Hitherto, there are many researches to investigate the structure under obtained data and analyze such data. H. Tanaka et al. proposed a Fuzzy regression model to illustrate the potential possibilities inherent in the target system. J. C. Bezdek proposed a Switching Regression Model based on a Fuzzy Clustering Model to formulate a forecasting model. The model proposed by Bezdek is to separate mixed samples coming from plural latent systems and apply each regression model to the group of samples coming from each system. So there are many achievements. However, it is hard to illustrate a rough and a moderate possibility of the target system. In this paper, in order to deal with the possibility of a social system, I will propose a possibility forecasting model.

View full abstract

In this paper we propose the exponential fuzzy numbers in Analytical Hierarchy Process (AHP),employing the fuzzy concept. The objective was determined a method of deriving the weights of criteria and their alternatives. Next, the influence of fuzzy is examined by using the membership function.Specifically we use the exponential fuzzy numbers for the pairwise comparison scale. Same experiments with AHP were conducted using the pairwise comparison scale and the new pairwise comparison scale of the method was examined.

View full abstract

We have proposed the processing methods for the fuzzy interval data which including vagueness of observer. And the usefulness of these methods are investigated. In these methods, to consider the vagueness of observed data, the correction value is added to the result which is obtained by normal statistical method using the representative values of fuzzy interval data. And as a application method, we have proposed the processing method in model selection by AIC. But, in this proposed method, the correction value is calculated by using the process of maximum likelihood estimation. That is, the proposed method dose not correct AIC directly. So, in this research, we consider the effectiveness of direct correction value in AIC. And the effectiveness of this method is discussed in the model selection of normal distribution.

View full abstract

Logistics network design is one of the most important phases in supply chain management (SCM). Transportation planning (TP) is a well-known basic network model that can be generally defined as a problem to minimize a total delivery cost. But for some real-world applications, the TP model is often extended to satisfy other additional constraints or is performed in several stages. In addition, the concept of inventory is not included in a traditional TP model. Moreover, time concepts, such as a carrying costs in a certain period, are not treated. Even if it is a plant that belongs to the same company, delivery methods may be different for each product. These restrictions on this model profoundly affect the use of the TP model in the real world. In this paper, we formulate the two-stage transportation problem with inventory and exclusionary side constraints. In this model, one year is divided into several terms and the annual demands of delivery centers are satisfied for each term. This model includes the additional constraint in which simultaneous shipments between some plants are prohibited.

View full abstract

This article presents the fuzziness of TOC(Theory of Constraints) and related basic researches. Mainly, we treated with TOC Project scheduling (TOCPS) Problem with fuzzy processing time and due-date. First, the ambiguous processing time and due-date are formulated by fuzzy possibility function and membership function, respectively. Next, we calculate agreement index between Manager's completion time of task and Decision Maker's due-date for finding common opinion. The method of transforming agreement index to actual processing time is proposed, and validity of the method is discussed.

View full abstract

In this paper, three types of c-means clustering algorithms are investigated with conditionally positive definite kernel. One is based on hard c-means, and the others are on standard fuzzy $c$-means and entropy regularized one. First, based on that conditionally positive definite kernel describes a squared Euclid distance between data in feature space, these algorithms are indicated from revised optimization problems of the conventional kernel $c$-means. Next, based on the relationship between positive definite kernel and conditionally one, the revised dissimilarity using conditionally positive definite kernel between a datum and a cluster center in feature space is shown. Last, it is shown that conditionally positive definite kernel c-means algorithm and kernel c-means algorithm with positive definite kernel from conditionally one are essentially equal with each other. An explicit mapping for conditionally positive definite kernel is also described geometrically.

View full abstract

This paper proposes a method for the Kernel Learning Vector Quantization Clustering using an explicit mapping to the feature space, in contrast to the method of Learning Vector Quantization using kernel function so far does not use an explicit mapping.As a result the proposed method can represent cluster centers on the feature space.

View full abstract

Recently, in clustering which means a unsupervised classification method, the one with kernel function is remarkable because we can easily calculate inner products of data which are map from the pattern space to a high-dimensional feature space. Moreover, a clustering method with penalty vectors is proposed as of the methods to handle uncertain data. This method can naturally formulate uncertainty to optimization problem. In this paper, we propose a new clustering algorithm with penalty vectors and kernel function. The proposed method can calculate cluster centers and penalty vectors in feature space directly by using explicit mapping to the high-dimensional feature space. We use L2 and L1 regularization terms to introduce penalty vectors.

View full abstract

Clustering is one of the important tools to analize data structure, and fuzzy c-means (FCM) is a typical method of clustering. By the way, numerical data include some errors in many cases. It is difficult to classify such data. To solve such problems, Endo et al. have introduced the concept of tolerance and constructed the algorithm of fuzzy c-means for data with tolerance(FCMT). In those algorithms, constraints for tolerance play an important role. In this paper, we will try to remove the constraints by introducing the penalty term instead of those. Moreover, L1-norm based FCM has been constructed. We will try to compare L1-norm based method with L2-norm based method.

View full abstract

A linear fuzzy clustering method, which is an extended version of Fuzzy c-Medoids, was proposed in order to extract intrinsic local linear dependencies from relational data. In this research, the characteristic feature of the method with non-Euclidean relational data is discussed through application of data transformation used in non-Euclidean-type Fuzzy (NERF) c-Means. An experimental result demonstrates that the data transformation makes is possible to find a suitable set of medoids, even when the relational measure is not Euclidean.

View full abstract

Linear clustering is a promising technique for extracting intrinsic local linear sub-structures. This paper proposes a new approach for extracting cross-shape clusters formed by two lines, in which two lines identification is jointed into one process of finding a single linear sub-structure by rotating prototypes. By rotating one of the two lines around intersection, crossing prototypes are formulated only by a single basis vector. In order to solve the size problems caused by prototypical lines with infinite length, the combined objective function with FCM is also considered.

View full abstract

This paper proposes an approach for extracting square-shape clusters, in which the FCE algorithm is applied in conjunction with local coordinate rotation. By rotating data points around the centers of squares, four sides of prototypical squares are estimated as a single linear sub-structure. After the rotation phase, each linear sub-structure is extracted in a similar procedure with local minor component analysis. Because FCE also considers optimization of the FCM objective function, the proposed algorithm can estimate not only square-shape clusters but also grid-like clusters having long tails by changing the priority weight of FCM.

View full abstract

This paper discusses the application of the fuzzy c-means based classifier (FCMC) to large scale data sets. Large scale data sets contain a huge number of samples (patterns). The number can be reduced by sampling, but the accuracy of the classifier on the test set may deteriorate, and the accuracy on the available data worsens. The FCM classifier uses covariance matrices whose size does not increase with the number of training samples, and the training time is proportional to the number of samples. By comparing the performance of FCMC with the support vector machine (SVM) classifier, which is known as one of the highest performance classifiers, the paper shows that FCMC nearly attains the accuracy of SVM and surpasses it in the training time and the testing time.

View full abstract

The language specification of fuzzy systems description language (FDL) was settled on as a common standard of the system development language to implement the fuzzy theory. It is possible to develop the system that uses FDL on general UNIX-OS now. However, the main applied field of the fuzzy theory is a built-in control field. In a word, it is necessary to use FDL only in the built-in control field. So that, the language specification and implementation of FDL were suitable for the built-in system development was considered. As a result, we can correspond by the improvement of the language implementation even if we do not change the language specification. Moreover, we considered the computational effort and the memory consumption as a development guideline for the built-in system written in FDL. A concrete consideration is a usage of fixed zero point that takes the place of floating point, and LUT. In this thesis, we report on these improvement point and guideline.

View full abstract

In this study, the piecewise Lyapunov function based switching fuzzy controller is proposed to get a relaxed criterion for the H-infinity control of continuous T-S fuzzy systems. The particle swarm optimization (PSO) algorithm, which is useful for nonlinear optimization problem, is utilized with the LMI tool to get the optimal H-infinity performance due to that some conditions of the criterion are bilinear with the s-procedure parameters. The simulation results have shown the validity of the proposed H-infinity control method.

View full abstract

This paper presents design methods of sum-of-squares-based observers for polynomial fuzzy systems. SOS-based observer design conditions are derived for three classes of polynomial fuzzy systems. We show the utility of the derived design conditions through some design examples.

View full abstract

Recently, it has become frequent for us to change the PC environment by the version-up of OS, switching OS, changing the platform, and so on. In such an unfamiliar environment, it is sometimes difficult to select and launch the software applications in order to achieve the current task, not only for a novice user but also for an expert user. Moreover, in a PC environment in which quite a lot of software applications are installed, the user may lose oneself to retrieve the suited alternatives for the task, even if it is a familiar environment. In this study, we propose a method of software classification and visualization by using Self- Organizing Map. All of the software applications installed in a PC are classified based on their categories, features, and so on. The result of the classification is visualized on a window. The user can look over an entire set of software applications installed in the PC, and determine the feasible tasks by combining them. In this paper, we present the description of our method, implementation of user interface, and the experimental result of the evaluation of our method.

View full abstract

It is not necessary to decide the number of the clusters in advance at possibilistic clustering, , and the cluster center converges in the place where data crowded. Therefore, we have carried out some experiments to search crowed data using possibilistic clustering. The results show the promising feature of the proposed method.

View full abstract

Shell clustering methods partition data sets into several shell-shape clusters by extracting local circles or ellipses as prototypes of clusters. This paper proposes hard c-regression models (HCRMs) for shell clustering. HCRMs successfully detect ellipses by using the hard clustering approach and random initialization.

View full abstract