In this paper, we propose a music recommendation system based on a certain theme, entered by a user. Recently, the amount of digital music database has grown rapidly with the popularization of the internet. As a result, it has become more time consuming for users to find songs of their distinct taste or feelings at the moment. Now, many music retrieving and recommendation systems exist on the internet, but these conventional systems hold one or more of the following problems: (1) users cannot find songs if they do not know the specific title or artist of a song; (2) they cannot search music with more than one sensitivity word; (3) users can only choose certain keywords or sensitivity words from a list of words provided by the system. Hence, we propose a novel recommendation system that integrates semantic associative search, freeing users from these constraints, and enabling them to search for their desired music in their own words.

View full abstract

In this paper, a programming system for text-editing task that uses both natural language text and direct manipulation is proposed. The system is based on the idea of ""programming by paraphrasing"", which enables users with little knowledge of computer programming to make computer program by natural language texts. The case-based reasoning (CBR) method is used for paraphrasing so that the system can acquire the meanings of unknown texts from users. The direct manipulation method is also introduced into the system in order to improve the usability of the system, especially for novice users. The experimental results show that novice users can understand and use the system easily and that novice users can achieve the task as quickly as skilled ones, while novice users employ much more direct manipulations than skilled ones.

View full abstract

This paper describes a programming system with which even novice users can easily specify their complex requests as natural language texts, in other words, write programs in everyday language, to operate computing systems. Based on a study of the characteristics of task specification texts written by non-programmers, the processing model is designed so that implicit and ambiguous information in an input text is detected and transformed into an explicit executable program structure. Context dependency in programming is also considered. We have developed a prototype system that deals with personal email management tasks. In this paper, we explain our processing model, give evaluation results, and discuss its effectiveness and future work.

View full abstract

The smart help we propose based on application software manuals features four phases: (1) understanding the user's input text, which is a question about software operation; (2) matching the result of understanding with software manual text; (3) planning a dialog with the user; and (4) paraphrasing matching manual text to output for the user. Manual text found during matching is paraphrased using the analysis result found during understanding. In planning, the hetorical structure of the selected manual is used to determine which clauses in matching manual text are specifically relevant to the user's question. Four paraphrases for a novice user are identified in a subject experiment, and rules for these are applied based on linguistic features extracted during the understanding. We evaluated an implemented prototype, and discuss its applications and problems.

View full abstract

Kernelized clustering algorithms are successful to obtain nonlinear cluster boundaries. Among them, kernelized hard c-means based on cosine correlation is useful for the document classification. However, it has the drawback of a high computational effort when a kernel function is used. In this paper, we propose new time-efficient algorithms for kernelized hard c-means based on cosine correlation. Our approach is that on-line algorithms are kernelized instead of batch algorithms such as kernelized hard c-means. Numerical examples show the effectiveness of the proposed method.

View full abstract

The aim of the paper is to construct new fuzzy clustering which treat data with the tolerance on L1 space. We formulate the tolerance which means the permissible range of the error and introduce the tolerance into optimization problems which relate with clustering. The problems are solved. New algorithms are constructed based on the results of solving the problems.

View full abstract

A new clustering algorithm is proposed. This algorithm is based on fuzzy c-means for data with tolerance and use a nonlinear transformation from the original pattern space into higher dimensional feature space than the original one with kernel functions. The first, a clustering algorithm for data with tolerance are introduced based on entropy-based method. The second, an objective function in feature space is shown . The third, Karush-Kuhn-Tucker conditions of the objective function is considered. and this condition is re-expressed with kernel function as the representation of an inner product for mapping from original pattern space into higher dimensional feature space than the original one. The last, an iterative algorithms is proposed for the objective function.

View full abstract

This paper presents a fuzzy clustering based on kernel method. This method is an extension of a self-organized fuzzy clustering method, in which we define the self-organized dissimilarity using the result of fuzzy clustering in such a way that the dissimilarity of objects is influenced by dissimilarity of classification structures corresponding to the objects. The dissimilarity of classification structures is used for the weights of the dissimilarity of objects. Since there are various relationships between the dissimilarity of classification structures and the dissimilarity of objects, we need a variety of the weighting. This is one attempt at assuming the inner product in higher dimension space as the relationship between the weights and the dissimilarity using the kernel method. Several numerical examples show the useful features of the proposed method.

View full abstract

Recently, the document data of news, websites and papers have accumulated in the database every year. And the amount of information increases substantially. Therefore, we can not understand and search the matter from all data. In this paper, we proposed a new clustering model for information retrieval using kernel aggolomerative hierarchical clustering(KAHC). Moreover, we show a new visualization technique for the result of data using KAHC and kernel Fisher discriminant analysis(Kernel-FDA). This visually will be able to verify information on related to the result of the information data. As a result, the data that was classified on the feature space can be visualized on the low dimentional space. Through some examples, we show to display the retrieval result more effectively becomes possible.

View full abstract

To realize a reliable classification of data, a new clustering method by the K-Means algorithm using a split and merge procedure is proposed. After ordinary clustering by the K-Means algorithm, a method determining whether or not each decision region should be split is introduced. After splitting each target decision region into subregions by using the K-Means algorithm again, the other subregions except one region are merged into appropriate adjacent decision regions. The goodness of this procedure is demonstrated under some classification experiments.

View full abstract

Performance of control and diagnostic applications depend strongly on the accuracy of sensor readings. However, sensors used in these applications are not always functional and/or may be mis-calibrated. Detecting, and recovering from, sensor faults is not trivial, however, although a number of recent studies have shown success in addressing this problem via computational methods (as opposed to using multiple redundant sensors.) This seminar presents the application of several alternative methodologies from Computational Intelligence to the problem of sensor fault diagnosis and recovery. These methods include Auto-Associative Neural Networks (AANN), Fuzzy Logic (FL) based clustering as well as Artificial Immune Networks (AIN). The application of these methods in engineering and in particular in energy management and control systems, among other domains, is discussed and relevant conclusions are drawn. In particular we consider the merits and potential drawbacks of each of the methods and suggest how they may be suitably applied to appropriate application domains. The seminar concludes with an outline of research issues in sensor diagnostics and potential future directions in this field.

View full abstract

A knowledge-based genetic algorithm for path planning of mobile robots is proposed. Problem-specific knowledge and heuristic knowledge are incorporated into encoding, evaluation and genetic operators of the genetic algorithm. An evaluation method is designed according to the environment and path representation, and specifically aimed at easily evolving better solutions with specialized genetic operators. The specialized genetic operators incorporate domain knowledge, which use a small-scale local search based on heuristic knowledge, and play a crucial role in evolving feasible and good quality solutions. The knowledge-based genetic algorithm is effective and efficient for mobile robot path planning in complex static environments including clustered unstructured environments and complicated structured environments, and dynamic environments with suddenly appearing obstacles, moving obstacles and moving targets. By considering other mobile robots as moving obstacles, the algorithm can be applied to real-time multi-robot path planning applications. The effectiveness and efficiency of the genetic algorithm are demonstrated by simulation studies.

View full abstract

The offshore petroleum industry has been faced with new challenges for petroleum production in deep and ultradeep waters. Sometimes, the extreme water depth involves automated and remote control system for underwater deepsea equipment involved in petroleum production and processing plants. More recently, the subsea separation process plant for produced oil and gas in deepwater has shown to be viable for new oil discoveries in deepwater and also to increase oil production from mature offshore petroleum fields. The present paper initially gives an overview of developments in intelligent system design for offshore petroleum production and intelligent supervisory and control system for subsea oil and gas separation plants. This paper also presents an intelligent hybrid system to model and control one of the drilling bit performance evaluators, the rate of penetration.

View full abstract

Skin rejuvenation has attracted much attention in the field of dermatology. The simultaneous irradiation of pulsed light and radio wave is a promising one of the anti-aging treatments. In this paper, an evaluation method of the effectiveness of skin anti-aging treatments based on mathematical morphological image processing is proposed. The arrangement of collagen in the dermis is analyzed by the size distribution method based on mathematical morphology. The results indicate that the simultaneous irradiation is more effective than the irradiation of pulsed light only.

View full abstract

Texture often affects visual impression more than equal to the shape or the color. Therefore, we have been interested in the effect of the texture and have been investigating. We have studied a relationship between the visual impressions of black fabrics and their image features from the viewpoint of Kansei engineering, and the results showed that visual impressions of fabrics often respond to not only surface reflectivity or mechanical characteristics but also the complex micro-components and global structures of the image features of textures. In order to evolve these researches to wide use, we have been studying precisely what about of an arbitrary texture influences human visual impressions and sensibility. In this paper, texture was altered and transformed by using the parameter estimation method of the texture based on Mathematical Morphology as a new step, and evaluated the relation between the parameter and visual impressions.

View full abstract

A method of optimizing structuring elements of morphological filtering for noise removal in texture images without any example of noiseless images is proposed. This method is based on the property that texture images are composed of repetitive appearance of microstructures, and the method applies the estimated shape of the microstructure as the optimal structuring element of the morphological opening and closing. The robustness of the estimation under the existence of noises is improved by introducing limitations to the pixel value distribution on the structuring element in the process of optimization, and iterating the optimization with modifying the size of the structuring element.

View full abstract

By using various sampling windows, morphological wavelets are generalized in order for efficient image compression and video coding. In the case of morphological wavelets which are based on max-plus algebra, the approximation signal is corresponded to the minimum (or maximum) element among the elements of sampling window, and the detailed signals are corresponded to the subtraction of the fixed element and other ones of the sampling window. Therefore, the proposed wavelets can be generalized without any restriction with respect to the sampling window. Through the coding experiments by using test images/videos extracted from standard image/video database, the performance and properties of the morphological wavelets by using 5 different sampling windows are shown.

View full abstract

From a temporal logic, by eliminating formulas containing additional pair of modal operators, one obtains a 1-dimensional normal modal logic. The temporalizaion problem is the inverse problem of this. Morphological dilation and erosion operators correspond to modal operators: a dual pair to a usual modal pair and adjoint pair to a part of temporal quadruple. Morphological Analysis makes it clear the relation between ""duality"" and ""adjunction"" of morphological/modal operators. By using this, we solve the existence problem. Namely, for any normal modal logic, there exists a temporal logic such that by elimination, the original modal logic is obtained.

View full abstract

We could generally analyze the inexact information efficiently and investigate the fuzzy relation by applying the fuzzy graph theory. We would extend the fuzzy graph theory, and firstly propose a fuzzy node fuzzy graph. Secondly, we transform it to a fuzzy graph by using T-norm family. In this paper, we would discuss about four subjects, (1) fuzzy node fuzzy graph, (2) new T-norm ""quasi logical product"", (3) decision analysis of the optimal level in a partition tree, (4) decision analysis of the optimal fuzzy graph in the fuzzy graph sequence . By using the fuzzy node fuzzy graph theory and this new T-norm, we could clarify the relational structure of fuzzy information, and by using the decision of an optimal level on a partition tree, we could analyze the clustering relation among nodes. Moreover, we would illustrate its practical effectiveness with the case study concerning sociometry analysis.

View full abstract

As for analysis of inexact information such as human behavior, mental process, social structure and so on, we consider fuzzy graph of some relation in these problems. Fuzzy graph is constructed with clusters of some levels. Furthermore, recently Uesu and Shinkai et. al [6] - [8] proposed a rational method to decide the optimal level of fuzzy clustering with partition tree. In this paper we improve their methods to use AIC (Akaike's information criterion) which is a likelihood estimator via Kulback-Leibler information number.

View full abstract

We often use fuzzy graph to analyze inexact information such as instruction/cognition sequence, sociometric structure, opinion poll and so on. Concerning the cluster analysis of a fuzzy graph, it is interesting how to decide the optimal cutting level as to a partition tree. In this paper, we would not only describe the decision method, but also discuss the approximate analysis of fuzzy graph. Moreover, we would illustrate practical effectiveness through the sociometry analysis.

View full abstract

It often happens with TVs or computers that a person would watch the display screen while lying on the side. This study describes the experiment concerning the effects of the relative angle between the body and display screen on the legibility.

View full abstract

Abstract The Hybrid GA combined with the inversion is effective for the TSP because the inversion is useful for solving the twist of the gene. On the other hand, the inversion in a simple GA (sGA) cannot show a high performance ex-cept in special cases such as a long size chromosome. Thus the inversion has rarely been used. Based on such a background, this paper tries to improve the convergence characteristics of an inversion. In the inver-sion, we treat several individuals, and process these individuals sequentially. The inversion leads to improved convergence characteristics compared with three types of crossover: normal inversion, Maximal Preservative Crossover (MPX), and Improved Edge Recombination Crossover (IERX), through the automatic fuzzy graph drawing.

View full abstract

We propose a object imaging by considering ultrasonic refraction. We do an experiment by using acrylic boards to imitate the skull bone and a steel with ditch to imitate cerebral sulcus. We first calculate the thickness of acrylic boards by using fuzzy logic. We next calculate the refractive angle of ultrasonic wave and correct the refraction of ultrasonic wave. In the result of applying this method, can determine width of the steel ditch at 5.57% of mean error ratio, and depth of the steel ditch at 3.39% of mean error ratio.

View full abstract

This paper proposes new determination method of the bone axis of the femur by integrating an ultrasound device with a 3D electromagnetic device. Recently, quantitative measurement method intra-surgery has been received grate attention. Problems how to determine the bone axes are always raised in the clinical situation. Especially, the conventional method usually uses the major trochanter as a reference point, however, it cannot be access in anterior cruciate ligament reconstruction. This study provides one of the solutions about the problem by integrating an ultrasound device with a 3D electromagnetic device. In the experiment, our method was compared with the conventional method where both of methods were applied for one cadaveric body. As a result, the 6DOF calculated by our method without using the major trochanter, was similar to the conventional method.

View full abstract

We have developed extracting the inner and outer walls of the cardiac muscle from a set of ultrasonic B-mode images. In this paper, we propose a new 3-D (radius, angle and time) display method of the motion function on the inner wall to evaluate the motion of the myocardium objectively. Its method displays the time transition of a cardiac wall during one cardiac cycle, and it is colored the thickness or the thickening rate of the myocardium on the surface of the 3-D object. The inner and outer walls of the heart are extracted from a set of ultrasonic B-mode Images. We implemented 3-D display system using C language and DirectX 3D in MS-Windows. We calculate the position of the points for 3-D display from the contour lines of the cardiac muscle by the cubic spline curve. The curve connects smoothly between one point and the other one point. And we found the thickness and thickening rate by using the contour lines of the cardiac muscle. We define the thickness by the shortest distance between the outer wall and the inner wall, and we calculate the thickness rate from the obtained thickness. We add the rotation view of the 3-D object to see the every side of myocardium. We confirmed the usefulness of our 3-D display.

View full abstract

A design idea of the Life support system which is expanded from the EMR (Electric Medical Record) system which has already been discussed and developed on the Linux base is considered. And it is described that as a communication system, the UBITSUKYI which is system name was developed on the basis of the Finger Braille technique. And from the point of support of handicapped persons a design idea of the Life support system into which the UBITSUKYI is embedded as an expanded module added to the EMR system is added and developed. After that, the essential meaning of the ubiquitous system is considered and the application of it is discussed by considering of the negative side of the today's ubiquitous system and some important abilities and possibilities of the Life support system are estimated.

View full abstract

In this paper we present a fuzzy neural network system for diagnosing osteoporosis based on mandibular cortical width and shape measurement on dental panoramic radiographs. The system consists in partitioning the input space into a set of subspaces using a novel fuzzy thresholding we propose in this paper. The fuzzy inference system constructed from the partitioning step is then incorporated with multi layer perceptron (MLP) neural network. The sensitivity and specificity averages are 90.92% and 63.69%, respectively.

View full abstract

This paper presents a new approach to perform maximum likelihood estimation in latent variable methods using Sequential Monte Carlo methods.

View full abstract

This paper attempts to perform on-line change detection given data sequence from unknown nonlinear systems. The proposed algorithm estimates the probability of occurrence of an abrupt change within a on-line Bayesian framework. The implementation is done via Sequential Monte Carlo (SMC). The proposed scheme is tested against two examples and evaluated by receiver operating characteristics (ROC) curve.

View full abstract

Recently there has been much interest of particle filters for nonlinear state estimation. A posterior probability distribution of the state variable is evaluated based on observations by particle filters in Monte Carlo simulation. However, almost actual systems have time varying properties. System model of some class of such systems consists of state dynamics and mode transition probabilities. In this paper, we consider state and mode estimation problem for stochastic hybrid systems. A multiple hypothesis filter is introduced to state and mode estimation, then hypotheses pruning based on the likelihood ratio is employed to reduce computational burden. Particle method is used for each hypothesis instead of conventional Kalman filter or extended Kalman filter. Numerical simulation result indicates the applicability of the proposed approach.

View full abstract

We report the first application of the ensemble Kalman filter (EnKF) to an intermediate coupled atmosphere-ocean model by Zebiak and Cane [1987] into which the sea surface height (SSH) anomaly observations by TOPEX/POSEIDON (T/P) altimetry are assimilated. Smoothed estimates of the 54,403 dimensional state are obtained from 1981 observational points with 2048 ensemble members. While data assimilated are SSH anomalies alone, an ensemble experiment of 2002/03 El Nino event based on the EnKF can predict consistent Nino 3 sea surface temperature (SST) anomalies about 5 months in advance.

View full abstract

Data assimilation in geophysics is a method of combining a numerical simulation model and incomplete observation data. A data assimilation problem can be formulated by an estimation using a nonlinear state space model and can be resolved using the particle filter. In this paper, we give a framework for resolving the correction and identification problem of erroneous tsunami simulation models using data assimilation. A numerical experiment is also conducted and its result shows the validity of applying data assimilation and the particle filter to a tsunami simulation model.

View full abstract

The theory of refinement in agglomerative clustering is studied. Although a simple relation holds as the necessary and sufficient condition so that a method generates refinements of clusters of the single linkage method, whether or not generalizations hold without the single linkage is unknown. Here a counterexample is shown between the average linkage and the complete linkage, which shows this theory is far from trivial. Observation of this theory leads us to a family of new methods using convex closures. Problems to be solved for the convex closure methods are discussed.

View full abstract

In this paper we study the use of fuzzy clustering for its application to classification problems. We apply the approach to the classification of gene expressions. We discuss the problem of aggregating the fuzzy partitions obtained from the fuzzy clustering. We propose an algorithm similar to $k$-nearest neighbor for solving some of the difficulties that appear in this problem.

View full abstract

Independent Component Analysis (ICA) is a technique for blind source separation and is also useful in regression (prediction) task when only a subset of random variables is observed. Because the task of regression by ICA consists of finding non-Gaussian components, it is closely related to projection pursuit regression. Local independent component analysis (Local ICA) is a non-linear extension of linear ICA models that extract local feature values by applying linear ICA in conjunction with suitable clustering algorithms. This paper proposes a switching regression model, in which local linear structure is first captured by fuzzy clustering, and then a non-linear regression model is estimated by a modified ICA model considering fuzzy memberships in each cluster.

View full abstract

A postsupervised FCMclassifier using a modified Cauchy weight as a membership function has been proposed. The performances of the classifier in terms of ROC and rejection curves are compared by using the benchmark datasets of the UCI ML repository. Cauchy weight function gives lower confidence to the classification decision and this distinguishes the FCM classifier from the Gaussian classifier. The FCM classifier outperforms the Gaussian classifier and the k-nearest neighbor classifier on many datasets in general.

View full abstract

Table-based collaboration support systems are an emerging technology that uses advantageous digital information technologies to enhance a traditional table, which is a well suited workspace for the collaboration of multiple workers. The technical advantages of such a table-based system (1) enable participants to naturally share non-verbal modalities such as eye contact and gestures and (2) allow users to hold discussions by using both real objects and information displayed on a horizontal screen. Various researches about table-based systems are in progress. This paper overviews such systems by focusing on the three major issues among them. The first issue is an interaction method suitable for table-based systems. Multi-touch input methods have been proposed to simultaneously enable participants to manipulate displayed information. The second is the orientation of displayed information. Table-based systems suffer a common problem: not equally recognizing shared information such as text and images because the orientation of their view angle is unfavorable. Solving/utilizing such inequality of viewing information is an important issue. The third is cooperated action between objects and tabletop surfaces. Methods have been proposed for providing haptic feedback to objects and utilizing objects as tiny screens. This paper also introduces two prototype systems as concrete examples of ongoing researches: the Lumisight Table and the t-Room. The former can simultaneously provide different images to users around the table. By using these characteristics, a multilingual face-to-face idea generation support system has been developed for the Lumisight Table. The latter is a work environment that aims to support collaboration executed between two distant places. Tabletop display in the t-Room is utilized as the basis for sharing physical objects between two distant places.

View full abstract

Dempster's rule of combination is the standard way of combining multiple pieces of evidence given by independent sources of information. However, it aroused controversy about the validity and many alternative combinations were proposed.The paper examines the model of combination discussed in the original paper, and indicates that the mathematical model is somewhat strange from the viewpoint of semantics. It might confuse the independence among information sources with the independence among bodies of evidence. The paper proposes a new model of combination. It introduces multiple belief spaces to represent multiple bodies of evidence, each of which is given by the independent sources of information. The bodies of evidence in the different belief spaces could be combined into another belief space possibly in three ways; by exclusion, by union and by compromise. Among them, the paper proposes the last one, combination by compromise. The combination was evaluated with some numerical examples that show the drawback and advantage of the Dempster's rule. The results show that the proposed is one of the most preferable combinations among alternatives.

View full abstract

This paper introduces new approaches to fuzzy inference system for system modeling based on input and output data using kernel machines. It is important issue how to select the best structure and parameters of fuzzy model from given input-output data. To solve this problem, this paper proposes the state-of-the-art kernel machine as the fuzzy inference engine. The kernel machine contains two modules such as the machine learning and the kernel function. The machine learning is a learning algorithm. The kernel function projects input data into high dimensional feature space. In this paper, the Support Vector Machine (SVM), Feature Vector Selection (FVS) and Relevance Vector Machine (RVM) as a kernel machine are presented. The proposed fuzzy system has the number of fuzzy rules and the parameter values of membership functions which are automatically generated. The structure of the proposed fuzzy inference system is equal to a Takagi-Sugeno fuzzy model. In addition, the number of fuzzy rules can be reduced under the condition of optimizing a given criterion by adjusting the linear transformation matrix or the parameter values of a kernel function using the Simulated results of the proposed technique are illustrated by examples involving benchmark nonlinear systems.

View full abstract

This paper describes a new practical autonomous lane navigation strategy for mobile robot that called ""Vision-based open loop control navigation"" strategy. The proposed ""Vision-basd open loop control navigation"" consists of two-phase of navigation process. In the first-phase, the mobile robot acquires image sequences and complete image processing iteration until enough confident level to generate appropriate path. In the subsequent, the mobile robot navigates based on appropriate path by applying open loop control manner. Separation of two-phase process can achieve reliable navigation in various types of outdoor environments. The validity of proposed strategy can be examined by actual experiments.

View full abstract

In this paper, we present a basic concept to design a Lyapunov function candidate and the stabilizing controllers of nonlinear discrete-time systems. First, we propose a stabilization algorithm for linear step-varying discrete-time system via piecewise Lyapunov functions based on a tracking trajectory control. The piecewise Lyapunov functions satisfying arbitrary trajectory control can be constructed. The design conditions are represented as LMI conditions. Therefore, the solutions can be solved efficiently. The algorithm is applied to nonlinear discrete-time systems. Finally, some illustrated examples are given to show the validity of the proposed method.

View full abstract

Human's action decision is based on wide targets and can respond flexibly under different situations. However, flexible adaptation to the situation with dynamic restriction is difficult but important in autonomous control system. In this paper, a method of developing intelligent controller based on soft target is proposed. Here, soft target is defined as a target set which includes many target elements with different satisfaction grade and is converted into target setting knowledge by soft computing. And this controller is applied to the control of inverted pendulum system with dynamic restriction to realize situation self-adaptation and target self-regulation in the process of swinging up and stabilization. The simulation results show the validity of the method.

View full abstract

Recently, there are a lot of research on swarm intelligence using multiple mobile robots. In most of the research, based on visual information of video cameras fixed to a ceiling, the mobile robots detect their own position. However, since the video cameras are fixed to a ceiling, there is a problem that the experimental field is limited. The goal of our research is to support the detection of the robots' position by using micro-helicopter with micro-camera to extend the experimental field freely. As the first step to the goal, this paper presents flight control of micro helicopter with coaxial counter-rotating blades. By employing fuzzy control approach to control the helicopter, we achieve hovering control and path-following control of the helicopter.

View full abstract

This paper describes a basic study on electroencephalogram (EEG)-based control. Tanaka has proposed the discrimination methods for detecting left/right direction thinking using brainwaves, in which the purpose is to improve recognition rates. Then the effects of the measurement position on the human head and the relationship with brain sites are not investigated. This paper employs Linear Discriminant Analysis (LDA) for this discrimination of human thinking. LDA enables us to investigate input / output relationship such as efficacy of each measurement position and relationship with brain sites. The authors focus on alpha and beta waves which are specific frequency components included in brainwaves, and we apply LDA to extracted alpha and beta waves. Experimental result for the data of brainwaves of 4 subjects analyzed by this method shows that average discrimination rate is approximately 70%. This paper also shows the investigation for the effective measurement position and the relationship between the thinking and the brain sites.

View full abstract

Visual perception-based autonomous virtual agent in virtual reality as a counterpart of an actual robot movement with a given dynamics is investigated. The visual perception is mathematically modelled as a probabilistic process obtaining and interpreting visual data from an environment. The perception obtained in the form of measurements in 2D is used for perceptual robot navigation. The visual data is processed in a multiresolutional form via wavelet transform and optimally estimated via extended Kalman filtering in each resolution level and the outcomes are fused for improved estimation of the trajectory. The computer experiments are carried out with perception measurement data, and the sensor/data fusion experiments are carried out by means of simulation in virtual reality. The improvement on the trajectory estimation by means of sensor/data fusion is demonstrated.

View full abstract

n this paper, we propose a novel framework for representing ""design patterns."" They are configurations of artificial environments surrounding people. Our framework extracts ecological factors of artificial environments from design examples. Focusing on indoor spatial designs and using the traditional hill climbing algorithm, we sample data on physical appearances of a certain architectural space. The data are organized into some clusters that are translated into Constraint-Interval Fuzzy Set (CoIFS). In terms of CoIFS, we can implement human-centered measurements. As a whole, the design pattern is represented by CoIFSs and contiguity among them. We discuss on the validity of our framework with a computational simulation. Extracted design patterns are instantiated into environmental factors of another design.

View full abstract

This paper proposes an intelligent control of multiple partner robot for interaction with humans. First of all, we discuss the current state of researches on the multiagent systems. We focus on the communication between multiple partner robots and multiple humans. If several human groups exists, the robots also should be divided into several groups according to the state of human groups. In this paper, we propose an intelligent control method for partner robots and their clustering method. Finally, we discuss the effectiveness of the proposed method through several computer simulation results.

View full abstract

This paper investigates role of humming sound in the healing process by Bhramari Pranayama (BP) which has been recommended, in many literatures on yoga, for removing stress and healing many mental abnormalities and has also been claimed to be recuperative for the same by many practitioners. Pranayama is a Sanskrit word which literally means a yogic act performed for controlling flow of vital energy in the body and mind. One of the main acts during doing BP is that subject imitates humming sound of bumble bee while exhaling. The analysis of the humming sound produced during pranayama shows a little variation in pitch confined within lower frequencies with concentration of energy in the lower frequency bands. Nasal formants are also present. The humming sound has also vibrato. It is hypothesized that such sound produces not only bone conducted audition but also vibration in cranium and changes in the brainwave patterns leading to relaxation and healings.

View full abstract

In this paper, a network distributed monitoring system was developed to improve the interaction among the users and local service robotic system and enable a remote user to get a better understanding of what is going on in the local environment. Network monitoring system using QuickCam Orbit cameras was developed to visualize the robotic system working, and was demonstrated from June 9 to June 19 at the 2005 World Exposition, Aichi, Japan. Improvements of network distributed monitoring system have been done from practical viewpoint. This paper presents the architecture of the developed network distributed monitoring system for monitoring the state of robotic system working and the state of the aged or disabled. Robot Technology Middleware (RTM) was used in the developed system. RTM is based on CORBA (Common Object Request Broker Architecture), so it has high scaling and inter-operating ability and facilitate network-distributed software and sharing. By using RTM, we can develop cameras functional elements as ""RT software component"", which makes application and system integration easier.

View full abstract