The techniques of artificial intelligence (AI) have been rapidly developing in video games. One of the main aims of AI implementation is to win against the human players. This is already achieved in some games, such as chess and shogi. Nowadays, the interest is moved to enhance the enjoyment of the players. Many researches are devoted to develop a human-like AI player to entertain human player. In this paper, we focused on a video game named PuyoPuyo. PuyoPuyo is one of the tile-matchins puzzle game controlling two pair of falling blocks. Random distribution of the blocks makes it difficult to set them into the ideal pattern. Skilled player seems to have some ideal pattern template models in their mind. This paper tried to derive these templates from the placements of blocks stacked by the skilled human player. Using these templates, it is partially capable of reconstruction the player's placements.

View full abstract

In this study, we aim to acquire a flexible ability of decision making for a robot. We are developing a cleanup procedure estimation system based on environmental situation that is measured moving objects and people. First, we developed an environmental measurement and management system that measures objects and people movement.However, measured data should integrate to significant data because the measured data differ in measured timing and accuracy. This paper explains a measurement and management server that is easy to use for deciding action of an interaction robot.

View full abstract

We have a chemistry of personality because we have communicated to the other by using own communication style. In this paper, we developed lifelike agents with some personalities and validated whether there is a chemistry of personality between human and agent. As the results it was shown there is an another chemistry of personality between human and agent by the experiments

View full abstract

Memory on human conduct intellectual creative activity is important. However, interaction with advanced information media is complicated, in a memory by using the complex interactions that an adverse effect is a problem. There are two types of short-term memory and long-term memory in the storage, there is a need to examine the issue from both sides. As a matter of short-term memory, even though the capacity of short-term memory is limited, forgetting of ideas it occurs and will use the capacity of short-term memory in the interaction. As a matter of long-term memory, due to the lack of meta memory of storage that rely on external storage devices, dead storage of memory occurs it can not access the storage. In this study, we pointed out the problems with the interaction of the information media, we propose solutions.

View full abstract

Research on physical human-robot interaction hasbeen attracting attention recently, focusing on robot embodiment. The work reported here proposes Active Touch Communication Robot(AcToR), a robot that is modeled on the hearing dog. A hearing dog is a type of dog assist people who are deaf or hard of hearing by alerting their handler to important sounds. AcToR uses the sense of touch to notify a human of the intention to transfer information. For example, when AcToR detects that a cell phone that is in another location has received a call, AcToR moves to the user's location and makes contact with the user's body to notify the user of the incoming call. The AcToR robot is based on the Roomba and uses the Roomba's bumper and contact sensors to detect contact. This paper reports the results of psychological experiments using the AcToR robot that indicate the feasibility of using touch to transfer information from a robot to a person.

View full abstract

In facing the elderly people society problems in Japan, the needs of robot partner as one solution is increasing. However, in order to realize natural communication and recommend/persuade elderly people, understandable or affective robot partner development is going to be needed in the process. Therefore, this paper discuss about affective robot partner for realizing natural communication with elderly people.

View full abstract

In our laboratory, the Interactive Chaotic Amusement System (ICAS) using network of chaotic elements have been already developed. We have also studied the calculation method of relaxation degree from the frequency power content ratio and the evaluation of human impression for the relaxation effect of generated sounds. In this research, we try to construct the system to judge the degree of relaxation based on the heart rate variability and to adjust the automatic operation to obtain the relaxation effect by presenting the sound to the user. In this paper, as an advance research on the construction of this system, we analyzed various features of the heartbeat (electrocardiogram) measured in a different state and found an index that judges the relaxation degree by using fuzzy inference. And we performed the experiment to generate a relaxation sound so that the subject might be felt relaxation adopting interactive sound generation system ICAS.

View full abstract

Recently, the growth of smart devices became able for us to easily use the personal data stored in smart devices. In this paper, we propose a method of using personal data for natural communication of robot partners with people. Next, we discuss the effectiveness of the proposed method through several experimental results.

View full abstract

Body motion expressed by onomatopoeias promotes tacit knowledge intuitively. On the other hand, individuality and ambiguity onomatopoeias have may cause misunderstanding in sharing body motion. The study calculated the similarity of writing-skill features expressed by onomatopoeias based on multi-dimensional fuzzy sets.

View full abstract

In recent years, development of the study support application using a tablet device is performed briskly. However, the application of existing is limited use. It is not easy for the user of the application to edit button and image freely. In addition, it is necessary for a user to choose application to a purpose because the learning content of the application is decided. This study is development of the application that a teacher can make the teaching materials in accord with personal ability. It can register a text and a voice into the arbitrary places of the picture. You can make educational materials for example calculation problems, English word problem and communication tool using this application.

View full abstract

Recent physical strength test have demonstrated that modern children have decreased physical strength level. The decline in physical activity in daily life is considered as one of the reasons. Children with disability compared with normal children, in addition to low fitness levels. So, establishment of a supporting method to keep the health of children with disability are sought as soon as possible. So developed systems that supports exercise for children with disability in this study. The aim of this system is two. First one is to increase the opportunities for children with disability to exercise by use of the system, and the second one is to strengthen the motivation to exercise for children with disability. As a result, be expected to establish their exercise habits. This system is measure the exercise using a Smartphone, and save the data to the cloud server. The saved data can be see through the browser. Teachers and parents use this data to support exercise for children with disability.

View full abstract

Since we can use calligraphy characters fonts on a computer, it is easy for anyone to write calligraphy characters without a brush. But it is impossible to express features of calligraphy characters particularly such as scratched look or blurred look because one calligraphy characters font has one shape of calligraphy characters. A prior research solved this problem and made calligraphy characters fonts which is scratched or blurred, but the system another problem. We can't adjust scratched or blurred look of the fonts by the system. Then, in my research, we make a system which we can use easily by innovating onomatopoeia.

View full abstract

In recent years, expression education that improves imagination and communication skills of learners has been attracting attention. Current education is carried out in a one-to-many situation. Therefore it is not clear whether each learner is able to achieve his/her objective in the education particularly for expression education, because it requires class situation that a teacher has one-on-one communication with a learner. In this paper, we propose a robot to support the expression education under one-to-one situation. We conducted an experiment to verify the effect of the robot. The result showed that the robot has same effect as when performing the expression education with a human teacher.

View full abstract

Clustering gene expression data has been done by many researchers. They are using traditional methods of agglomerative hierarchical clustering, which fail to use domain knowledge obtained from vast gene studies. An example of such domain knowledge is gene ontology terms for messenger RNA. We propose two methods of clustering gene expression data and different experiments using gene ontology terms. One is traditional agglomerative hierarchical clustering using similarities based on gene ontology terms. Another is algorithms using the modularity index applied to networks including gene ontology terms. Clustering results of real large-scale experiments are shown.

View full abstract

Recently, in assessing harmful chemicals, low cost cell experiment has attracted a lot of attention as an alternative to animal experiment.Also in acquisition of gene expression data, it is able to acquire the large data more accurately since the next-generation sequencer is implemented. However, methods for analyzing gene expression data acquired by next-generation sequencer from cell experiment have not been established yet. In this study, we propose one of such method: a method to extract RNAs of characterizing a correlation of gene expression data. Specifically, Optimal problems for the correlation coefficients are given and the optimal solutions are acquired by the meta-heuristics. Information between gene expression data is summarized by the proposal method. The features of sets of RNAs are discussed and the effectiveness of the proposal method is evaluated through data experiments.

View full abstract

Clustering user structures in SNS has been attracted interests of many researchers. Nevertheless, available clustering techniques are relatively poor, and there is much room for further studies. The aim of this paper is to propose new method based on extensions of traditional agglomerative hierarchical algorithms to directed graphs. The proposed methods moreover use core points studied in DBSCAN. We also propose the use of core points in methods using the modularity index, a typical is Newman's method. Numerical experiments based on real data from SNS are shown.

View full abstract

Discrimination methods have been proposed by developing a discriminator used with data contains discrimination information. However, when the observed data is multivariate data the discrimination becomes more difficult. In order to solve this problem, a variable selection is important to improve the correct identification rate. In this paper, a new variable selection method is proposed through the principal components by using the relationship between discrimination information and principal components. Several numerical examples show a better performance of the proposed method.

View full abstract

k-means clustering (KM) algorithm, also called hard c-means clustering (HCM) algorithm, is a very powerful clustering algorithm, but it has a serious problem of strong initial value dependence. To decrease the dependence, Arthur and Vassilvitskii proposed an algorithm of k-means++ clustering (KM++) algorithm on 2007. By the way, there are many case that each object is allocated on an unit sphere, e.g. text clustering. Honik, Kober, and Buchta proposed spherical k-means clustering (SKM) algorithm to classify such objects on 2012. However, the algorithm also has the same problem of initial value dependence as KM. Therefore, this report discuss the following points: (1) the dissimilarity of SKM is extended to satisfy the triangle inequality, and (2) spherical k-means++ clustering (SKM++) algorithm which works well for the problem is proposed. The report shows that the effectiveness of SKM+ is theoretically guaranteed.

View full abstract

Clustering is a technique of unsupervised classification. The methods are classified into two types, one is hierarchical and the other is non-hierarchical. Fuzzy Non-metric Model (FNM) is a method of non-hierarchical clustering. FNM is very useful because Euclidean spaceis not required on data space and only similarity or dissimilarity between data is needed. However FNM cannot handle uncertain data, e.g. incomplete data, or data which have errors. In order to handle such data, concept of data with tolerance has been proposed. The methods using the concept can handle the uncertain data in the framework of optimization, e.g. Fuzzy $c$-means for data with Tolerance (FCM-T). In this report, first we will propose new clustering algorithm to apply the concept of tolerance to FNM. Second, we will show that the proposed algorithm handle incomplete data. Third, we will verify the effectiveness of the proposed algorithm in comparison with other ones for incomplete data through some numerical examples.

View full abstract

Hierarchical Linguistic based Clustering (HLC) is one of linguistic-based clustering algorithm.This method uses rule-groups which are not consistent.This study proposes gravity-like rule in HLC and shows that the rule improves clustering accuracy.

View full abstract

In this report, some possibilistic clustering methods are proposed based on noise clustering.The standard possibilistic clustering is identical to the standard noise clustering when the cluster number is one, whereas the entropy-regularized possibilistic clustering is different from the entropy-regularized noise clustering.Therefore, another entropy-regularized possibilistic clustering method is constructed from the entropy-regularized noise clustering with the cluster number one, along with proposing some variants.

View full abstract

The x-means divides a set of objects without determining the number of clusters by using iterative k-means and evaluation criteria.A series of cluster validity measures is also used in order to evaluate the clustering results and determine suitable number of clusters. We propose cluster validity measures based x-means by introducing cluster validity measures instead of information criteria.We moreover show the effectiveness of the proposed methodthrough numerical examples.

View full abstract

k-anonymization by fuzzy k-member clustering has been shown to be useful in privacy-preserving multivariate data analysis with lower information losses. In this research, the anonymization model is applied to eigen-face-based personal identification system and the applicability to crowd movement analysis with fuzzy k-anonymization of face images is discussed. While k-anonymization process makes it difficult to identify each person, the anonymized information is shown to be still useful in capturing crowd movement in large public facilities.

View full abstract

Fuzzy Co-clustering based on multinomial mixture concept was demonstrated to be useful in cooccurrence information analysis among object-item relations, but lacks exclusive nature in item partitions. Then, partition quality may be improved by emphasizing mutual differences among co-clusters. In this research, the conventional fuzzy co-clustering model is extended by introducing a partition penalty for selected items, and its influences on partition quality is discussed through several numerical experiments.

View full abstract

Although semi-supervised clustering is a promising approach for improving partition quality in large-scale data analysis, it is often difficult to utilize an enough amount of supervised objects. A virtual sample approach is a practical technique for improving classification quality in semi-supervised learning, in which additional virtual samples are generated by combining several supervised objects. In this research, the virtual sample approach is adopted in semi-supervised fuzzy co-clustering and its characteristics are demonstrated through several numerical experiments.

View full abstract

A prediction method of game results using kick information is studied in this paper. Although it is not possible to exactly know a strategy that an opponent team is taking, that strategy is well represented by how the players in the team kick during games. Extracted kicks such as passes and dribbles form a kickdistribution, which hopefully represent the strategy. A series of computational experiments are conducted in order to examine the performance of the proposed method.

View full abstract

In this paper, we discuss formation control of multiple mobile robots with co-leader. Our objective is to realize the robust control method for applying the multi-robot system to various situations such as maintenance of the formation shape, obstacle avoidance and adjusting to environmental changes.First, we apply a spring model for constructing the formation and maintaining the optimum position between the robots. Next, we introduce the obstacle avoidance method by using fuzzy control.Finally, we show several experiments in simulation and real environments in order to verify the effectiveness our proposed method.

View full abstract

This paper discusses the collective consciousness created by multi-robot communication. We apply PALRO to multi-modal communication to discuss the collective consciousness from the viewpoints of systems theory and relevance theory. Finally, we discuss the difference between humans and robots based on systems theory and relevance theory.

View full abstract

In this paper, we propose an emotion expressing model that uses an appropriate color for the feelings of the robot that emotion develops. In the conventional study of emotion expressing of robot, as an affective transmission method, robot's facial expression have been used. However, we can use it for only a robot having a face by the facial expression method. To solve this issue, we propose a method to express emotion by a color for the robot which did various appearance. In simulation results, we confirmed that this model dose a color suitable for the affective development of the robot.

View full abstract

Recently, aging society is progressing in Japan, and the use of electric wheelchairs and mobility scooters are needed. These vehicles are moved by wheels, and it is difficult to get over a step. Therefore, barrier-free design in all facilities has been desired for a long time. However, we have to spend much time and a great effort to do achieve the thing.We study the machinery of the wheels that moves by the wheels on the level ground, and get over a step with the wheel which is unfolded. And, this kind of wheels can get over a step with good moving efficiency. Moreover, we study about the proper control of the degree of the wheel which is unfolded.

View full abstract

In recent years,robots have been used in various scenes .It is useful to work in hazardous locations where human can not be activities. There are many cases interrupted communication in the accident. So the robot which can do autonomous movement is required. It is necessary that creating an environment map and selecting an optimum route. In this study,we try an autonomous mobile robot to recognize the position and posture of its body by using a depth sensor and rotary encoder ,and use Dijkstra's algorithm to have the shortest path plan to the goal of the robot.

View full abstract

The personality disorder should be contemporary social issue in terms of mental health. In the study, the fact finding survey for college students of adolescence is designed according to the personality disorder diagnosis of DSM-5. The characteristics of BPD, NPD and APD can be summarized empirically from the statistics analysis. The estimation models for the degrees of personality disorders are proposed with linear regression analysis as well as neural network approach. It is concluded that neural network model can provide the proper combination of BPD, NPD and APD.

View full abstract

The spread process of clean energy vehicle opposing to conventional vehicles is discussed in the study to realize low carbon society. The decision making action about the vehicle choice at the vehicle purchase is modeled. The choice of the clean energy vehicle is based on potential consciousness, and uses the questionnaire survey. The vehicle choice model is decision making structure with much variety, ant the application of the statistical model is difficult. Therefore, the decision tree technique to clarify decision making structure is used. The fuzzy decision tree technique to describe a judgement based on the fuzzy information about the clean energy vehicle is used. The decision making structure of the driver for the vehicle choice is clarified, and the examination of the spread policy of the clean energy vehicle is enabled.

View full abstract

Multi-agent simulation approach is suitable to describe the heterogeneity on social interaction. The choice of clean energy vehicle corresponding to the promoting policy related to economic incentive or innovation of vehicle is estimated with the multi-agent simulation. The proposed multi-agent simulation system consists of the vehicle choice model and the social interaction model. The vehicle choice is affected by social conformity effect in the model. The relation between agents like a social network of the real world is generated with a simple algorithm by small world network model. The time series changes of the number of the clean energy vehicle and the volume of greenhouse gas emission are estimated by proposed multi-agent simulation.

View full abstract

In the field of RoboCup Middle Size League, self-localization is important to estimate own position including goal and robot positions and to decide strategy.Basically, we estimate the self-position with the image information, the environment information and the field information.However, the above method cannot finish processing in real time, so it is difficult to perform the strategic behavior and cooperative operation.In this paper, we construct the self-localization method in real time by applying genetic algorithm.

View full abstract

In the field of RoboCup Middle Size League, Omni-directional Mobile Mechanism is very important for moving and dribbling.We improved conventional Omni-directional Mobile Mechanism for reducing vibration, and moving smoothly.Thus we performed performance test of Dead Reckoning vibration and error for new Omni-directional Mobile Mechanism.In this report, we introduce the new Omni-directional Mobile Mechanism and as a result of performance test.

View full abstract

In recent years, from agriculture away from the like of the aging population and young people, it has been a problem that the proportion of farmers has decreased. From this, although activities of the robot is expected in agriculture, when operating a robot in a real environment, it is considered that many of the problems occur. So, this research a pre-stage of the actual introduction into the field, we have developed a tomato harvesting robot toward the tomato robot competitions. In this paper we describe the development of the tomato harvest robot toward the tomato robot competitions.

View full abstract

Accurate speed control is one of the key issues to successful application of mobile robot systems. In this paper, we describe a development of sliding mode based speed tacking controller for mobile robot. In order to design sliding mode surface, we identify transfer function of the mobile robot. Validity of developed sliding mode based speed tracking controller is confirmed both simulation and actual mobile robot.

View full abstract

Robot technology is very useful for learning task at the engineering education in the field of electronics and mechatronics, because it is consolidation of various engineering factors. Moreover, encouraging and maintaining the motivation is important in engineering education, because it affects the effect of education. In this research, we aim at developing the educational tool and verifying the effect for educational tool using comic. We report the verifying result of effect of developed educational tool.

View full abstract

In information engineering education, there are 2 fields, the softwere technique as represented by programing and the hardware technique as represented by computer architecture. At first, students learn each technique individually in National Institute of Technology , Fukui College. And the system development excerise involving these technique is arranged at higher grage to encourage students to understand fundamentals. However, it is difficult to make students conscious of the application of each technique. Authors have held the exersice for low grade students using microcomputers as a kind of the application to educate fundamentals of wide range of information engineering. We aimed the effect for learning motivation by experiences of application in eary grade.

View full abstract

Most of commercialized robots are developed for industry and those for agriculture are still on developing phase. Tomato is one of most produced vegetables and grown up in special facilities, however, the hot temperature, humidity and long woking time make the working condition harder and huge load. In this paper, we describe the tomato harvester robot through tomato robot competition.

View full abstract

The scale of the game machine market in recent years is decreasing a great deal after the Lehman shock in 2008, under having a peak previously in 2007. The world great three game makers, i.e., Sony (SCE), Nintendo and Microsoft are facing hard problems. It may be because of the continuous expansion of the game market with the use of the PC and the rapid prevailing of smartphones. In this paper, a questionnaire investigation is executed in order to clarify the current usage condition, issues and desired function etc. Fundamental statistical analysis, Factor Analysis, Multiple Correspondence Analysis and Quantification Methods were executed. Reviewing past researches, there are some related papers, but they do not include new machine tools which are increasingly used in recent years. This research presents up-to-date research results, which can rarely be seen in other researches. Some interesting results were obtained.

View full abstract

How to allocate goods in shop shelves makes great influence to sales amount. Searching best fit allocation of goods to shelves is a kind of combinatorial problem. This becomes a problem of integer programming and utilizing genetic algorithm may be an effective method. Reviewing past researches, there are few researches made on this. Formerly, we have presented a papers concerning optimization in allocating goods to shop shelves utilizing genetic algorithm. In those papers, the problem that goods were not allowed to allocate in multiple shelves and the problem that goods were allowed to allocate in multiple shelves were pursued. In this paper, we examine the problem that allows goods to be allocated in multiple shelves and introduce the concept of sales profits and sales probabilities. Optimization in allocating goods to shop shelves is investigated. Expansion of shelf position is executed. Utilizing genetic algorithm, optimum solution is pursued and verified by a numerical example. Various patterns of problems must be examined hereafter.

View full abstract

In industry, making a correct forecasting is inevitable. If the correct forecasting is not executed, there arise a lot of stocks and/or it also causes lack of goods. Time series analysis, neural networks and other methods are applied to this problem. There are some related researches made on this. However, it can be said that an application to sales forecasting is rather a few s a whole, In this paper, neural network is applied and Multilayer perceptron Algorithm is newly developed. The method is applied to the food production data of prepared frozen foods. When there is a big change of the data, the neural networks cannot learn the past data properly, therefore we have devised a new method to cope with this. Repeating the data into plural section, smooth change is established and we could make a neural network learn more smoothly.

View full abstract

How to allocate goods in shop shelves makes great influence to sales amount. Searching best fit allocation of goods to shelves is a kind of combinatorial problem. This becomes a problem of integer programming and utilizing genetic algorithm may be an effective method. Reviewing past researches, there are few researches made on this. Formerly, we have presented a papers concerning optimization in allocating goods to shop shelves utilizing genetic algorithm. In those papers, the problem that goods were not allowed to allocate in multiple shelves and the problem that goods were allowed to allocate in multiple shelves were pursued. In this paper, we examine the problem that allows goods to be allocated in multiple shelves and introduce the concept of sales profits and sales probabilities. Expansion of shelf is executed. Optimization in allocating goods to shop shelves is investigated. An application to the convenience store with POS sales data of cup noodles and cup soup is executed. Utilizing genetic algorithm, optimum solution is pursued and verified by a numerical example. Various patterns of problems must be examined hereafter.

View full abstract

In industries, how to improve forecasting accuracy such as sales, shipping is an important issue. There are many researches made on this. In this paper, a hybrid method is introduced and plural methods are compared. Focusing that the equation of exponential smoothing method(ESM) is equivalent to (1,1) order ARMA model equation, a new method of estimation of smoothing constant in exponential smoothing method is proposed before by us which satisfies minimum variance of forecasting error. Generally, smoothing constant is selected arbitrarily. But in this paper, we utilize above stated theoretical solution. Firstly, we make estimation of ARMA model parameter and then estimate smoothing constants. Thus theoretical solution is derived in a simple way and it may be utilized in various fields. Combining the trend removing method with this method, we aim to improve forecasting accuracy. Furthermore, "a day of the week index", is newly introduced to the daily shipping data of the sanitary materials and we obtained good result.

View full abstract

We discuss reuse of the knowledge which is extracted from the Q-table by FCM-type Co-clustering. Advantages and disadvantages of setting the initial values in Q-table based on the extracted cluster data and using the membership in updating of Q-table where several states are updated at same time are reviewed through the path search problem and the single pendulum standing task.

View full abstract

On-line dynamic fuzzy classifiers are investigated in this paper. Previous papers have proposed confidence-weighted learning method for modifying the consequent real values. The framework used in the previous proposal can be used in any type of learning if the value-modification is based on an mathematical optimization. This research focus on passive-aggressive learning and examine the performance of the proposed online dynamic fuzzy systems with the previous ones.

View full abstract

In Interval AHP, a comparison is considered to consist of the real values in the corresponding interval weights. The widths of the interval weights are minimized so as to be close to the given comparisons as much as possible. In the same concept of Interval AHP, this study proposes the model to obtain fuzzy weights as the extension of the interval weights. In order to obtain fuzzy weights, Interval AHP is modified by focusing on the lower bounds of the interval weights from the viewpoint of belief function in evidence theory. As the given comparisons are more consistent, it means that they are based on the higher-level sets of fuzzy weights, which are denoted as interval weights.

View full abstract

From the viewpoint that the vagueness of decision maker's evaluation causes the inconsistency of a pairwise comparison matrix, Interval AHP estimating interval priorities has been proposed. In this paper, we demonstrate the insufficiency of the conventional interval priority estimation method and propose several modified estimation methods which reflect more vagueness of decision maker's evaluation. In those proposed methods, interval priorities are estimated easily by solving a linear programming problem. Numerical experiments are conducted to compare estimation methods from the point of view of the accuracy in estimating interval priorities and dominance relations between alternatives. Through the experiments, the advantages of modified estimations over the conventional one are demonstrated. Furthermore, we investigate the best estimation method among them.

View full abstract