In this paper, we propose a novel Actor-Critic method in the agent environment and action space based on the normal Actor-Critic method and PSO. In this algorithm, particles are centers of some actions or some states, and roam through the space looking for appropriate segmentalized space. The purposes of this study are learning efficiency improvement and heuristic space segmentation. In our method, swarms move in the space during the agent's learning process. Appropriate segmentation can minimize the learning time and enables us to recognize the evolutionary process. Thus, this method is also designed for humanlike decisions in the learning process. The simulation results indicate that our method shows some clusters in the action and state space. Space segmentation, such as group formation, language systems and culture, will be revealed by multi-agent social simulation with our method.

抄録全体を表示

We have developed partical 2-legged robot, using spring and motors. This robot can be used to absorb the shock of landing, and utilize the energy of walking.It is important to choose the walking behavior to achieve this robot. In this paper, we analyzed the behavior of the robot using motion model and optimization algorithm.Moreover, we investigate walking control method.

抄録全体を表示

In general, legs and crawlers are used as a robot to climb stairs. But, driving performance of these robots in the plains is lower than the wheel robot. The robot of this research runs level ground like a wheel and gets over a step with the wheel which unfolded. In this paper, we introduce a expandable wheel system, and it's described how to control for wheel expansion on the stairs. The control method has to be optimized. It is the first report to which we did a basic analysis.

抄録全体を表示

The development of the space elevator attracts attention as a study of the precedence of the space development recently. We examine the elevator of the simulation to go up and down at several hundred meters above the ground as these technical fundamental researches now. We measure the posture of own in real time to operate an elevator for high speed / stability, and control of the posture becomes important. Therefore we introduced an attitude measurement system using accelerometers and gyroscopes. Then, we examined a method of the optimization for moving speed as the first step to perfect posture.

抄録全体を表示

This paper describes a system which generates operetta songs fitting to story scenes represented by texts and/or pictures. Inputs to the system are original theme music, numerical information on given story scenes and story texts. The system outputs operetta songs generated based on impressions of story scenes. The system generates variations on theme music and lyrics according to image of music and lyrics obtained from numerical information on given story scenes. Evolutionary computation is applied to generations of variations and lyrics. Using a vocal synthesizer and a general midi synthesizer, the system plays operetta songs as the variations on theme music with the lyrics. The system reflects user's Kansei to variations on theme music and lyrics using interactive evolutionary computation. This paper also describes the evaluation experiments to confirm whether the generated songs reflect impressions of story scenes appropriately or not.

抄録全体を表示

A purpose of this paper is to model a decision making process from the view point of Service engineering. We lead to consumer's change of internal state and lead to consumer buying behavior by Kansei information on the post card. In this experiment, we revealed consumer take a decision making based on the trust relationship with laundry owner. It means that to keep having a relationship with consumers by using information media leads to create a buying behavior of future and sales. Moreover, we got actual data of buying behavior in store experiment. In a series of these behaviors, we visualized consumer's decision making process by Text Mining. As a result, we revealed that consumers processed 6 factors, which are concerned with their daily life and modeled 6 factors decision-making process.

抄録全体を表示

The purpose of this paper is to identify, though qualitative research, the process by which people retain personal belongings for long periods by modeling the conceptual structure of how the possession of personal object is maintained. To collect the data, interviews were performed with people who had held valued possessions for long periods. The qualitative data in their comments were analyzed by M-GTA. The results show a process of updating the possession status of owned items by attaching meaning to them, This attachment of meaning happens as a result of the owners making decisions that are based on how they see the object's status while engaged in their daily actives.

抄録全体を表示

With the evolution and spread of information technologies including Web, the amount of information that users can obtain has dramatically increased. Huge quantities of information have made it difficult for users to reach their desired information. In order to address this problem, various information recommendation systems including cooperative filtering have studied and developed with certain results. Many of traditional studies on information recommendation provide a method to statistically determine, based on the user's preferences and behavioral history, the right information to recommend. Because the user's request depends on their situation as well, there are also information recommendation studies that take the user context into consideration. These traditional studies, however, aim at recommending highly satisfactory information in a one-time or small number of uses of a system or service, and never take it into account that the user would feel tired of the same system or service after they have used it many times over. In reality, if the user can obtain information to their preference, they would feel very satisfied with it at first, however, if they are only presented with similar information every time, they may eventually get tired of the system and cease to use it. Consequently, in order to realize an information recommendation system that the user wants to continue to use, you need consideration of a human nature of being tired of the same thing, as well as interaction design for information recommendation that leverages its characteristics. This study has modeled the human nature of being tired of the same thing and enabled the measurement of the degree of such boredom. Also, the study has implemented an information recommendation system that has features to (1) visualize the degree of the user's boredom and (2) recommend information that can reduce it. Then, an evaluation experiment was conducted using a system developed. The result shows that an information recommendation system proposed in this study can get the user less bored and more satisfied when used continuously. In addition, it enables the user to make various choices even in a limited content. These results suggest that both information recommendation based on a Kansei model of human boredom and interaction that visualizes the degree of the user's boredom and makes them conscious of it work well for their continuous use of the system.

抄録全体を表示

This paper proposes an anti-sway control scheme for overhead cranes with load hoisting by the use of evolutionary computation techniques. The overhead crane involves a planar motion in conjunction with a hoisting motion. In order to reduce the sway of the load after positioning, we design the trajectory for the position of the trolley. Radial basis function networks (RBFNs) are employed to generate the desired trolley position, and then a particle swarm optimization (PSO) is used as the learning algorithm, in which the maximum swing angle after positioning is adopted as the objective function. The suppression of the sway angle can be accomplished by moving the trolley along the generated trajectory. The ability of the proposed anti-sway control is confirmed by numerical simulations.

抄録全体を表示

This paper proposes a concept of "behavior simple" and "behavior composed". The behavior simple means a primitive behavior and the behavior composed is a combination of behaviors simple. An artificial creature firstly learns some behaviors simple. Then it learns the behavior composed as a combination of the behaviors simple. This concept is applied to an artificial flying creature (AFC). The AFC learns two kinds of flight, flapping and gliding, independently. After that, it learns a sophisticated behavior by alternatively selecting these behaviors. Simulation results prove the adequate flight by alternatively using flapping and gliding is obtained.

抄録全体を表示

Visualizing complex networks is a useful technology to acquire new knowledge since it allow us to overlook a whole network characteristic at a glance. In this paper, Self-Organizing Map (SOM) is adopted for visualizing the network. A conventional visualization method named Inversed-SOM (ISOM) brings us a great benefit for visualization, but it has disadvantages to disadvantages to distort the network unnaturally. We improve the ISOM to get rid of this disadvantage and propose a new visualization method named DSSOM (Dynamically-Signaling Self-Organizing Map). Numerical experiments prove that the DSSOM overcomes network distortion in visualization and that it is free from a scale size of visualization.

抄録全体を表示

In many E-commerce sites, recommender systems, which provide personalized recommendation from among a large number of items, are recently introduced. Collaborative filtering (CF) is one of the most successful algorithms which provide recommendations using ratings of users on items. There are two approaches such as user-based CF and item-based CF. Additionally a unifying method for user-based and item-based CF was proposed to improve the recommendation accuracy. The unifying approach uses a constant value as a weight parameter to unify both algorithms. However, because the optimal weight for unifying is actually different by the situation, the algorithm should estimate an appropriate weight dynamically, and should use it. In this research, we propose an approach for estimation of the appropriate weight based on collected ratings. Moreover, we discussed the effectiveness based on both multi-agent simulation and MovieLens dataset.

抄録全体を表示

This paper discusses genetic operations and their effects on evolution of chromosomes in BSIM parameter extraction problems. Generally, Real-coded GA using Simplex Crossover is often employed to extract BSIM model parameter sets. BSIM model parameters, however, have recommended operating ranges. There are regarded as a kind of constraint, thus all extracted parameters have to be satisfied these constraints. In many cases, when the number of parameters becomes large, the conventional method using GA and SPX generates a lot of infeasible solutions because SPX makes offspring on the simplex plane expanded by ε parameter. This reduces the effectiveness of GA drastically. Therefore, this paper proposes a new genetic operation considering these constraints to prevent reduction of effectiveness. In this paper, some experiments using actual static characteristic curves MOS-FET were conducted to validate the proposed method and the effectiveness of the proposed method is also discussed.

抄録全体を表示

We developed a dynamic production system of virtual flowerbeds using an interactive genetic algorithm on the Web. The initial population was consisted of 100 virtual flowerbeds. User selected about 10 flowerbeds as candidate parents. Next generation with 60 or 100 flowerbeds was produced by mating with each other in candidate parents and so on. Three ways of mating was used in this system, that is, 1) Two-point crossover, 2) Uniform crossover, 3) Crossover to preserve plant structure. 25 students used this system up to 7th generation. And the experiment of four-fold choices at a later date showed that the evolutional system produce more preferable flowerbeds than flowerbeds produced by manually. Furthermore, a MDS analysis using flowerbeds' hue components implied that there are no cluster structure and individual's preference is different respectively.

抄録全体を表示

Numerical optimization problems arise from every field of engineering, science, and business. To tackle the increasingly complex real world optimization problems, effective and efficient algorithms are always on active demand. However, the performance of many existing algorithms deteriorates rapidly as the dimensionality of the search space increases. One natural way to address this problem is to adopt the divide and conquer strategy. In its basic form, the algorithm partitions the high dimensional problem into low dimensional subproblems and optimizes them cooperatively. In this paper, we propose a cooperative particle swarm optimization algorithm for large scale numerical optimization. Firstly, a statistical learning model is proposed to assess the degree of interdependencies among variables. Secondly, based on the variable interdependencies, a method is proposed to decompose a large scale problem into overlapping small scale sub-problems. Finally, a cooperative particle swarm optimization framework is proposed to optimize the sub-problems cooperatively. To examine the performance of the proposed algorithm, experiments were conducted on 10 benchmark functions of diverse complexities. The experimental results show that the proposed algorithm is able to find the nearoptimal solution in most cases, which demonstrate effectiveness of the proposed algorithm.

抄録全体を表示

Recently MOEA/D (multi-objective evolutionary algorithm based on decomposition) was proposed by Zhang and Li (2007) for difficult multi-objective optimization problems. The main feature of MOEA/D is the decomposition of a multi-objective problem into a number of single-objective problems to simultaneously optimize a scalarizing function with different weight vectors. The point is the use of a neighborhood structure of weight vectors to optimize similar single-objective problems in a cooperative manner. Each weight vector has an elite solution with respect to the corresponding single-objective optimization problem. The diversity of solutions is maintained by the use of uniformly distributed weight vectors while the convergence of solutions is realized by the use of scalarizing function-based fitness evaluation. One important issue in the implementation of MOEA/D is the choice of an appropriate scalarizing function. In our former study, we proposed a simple idea of simultaneously using two different scalarizing functions. In this paper, we further examine this idea. That is, we discuss how to choose two scalarizing functions. We also examine the simultaneous use of three or more scalarizing functions.

抄録全体を表示

This paper presents a study on the localization correction for the mobile robot using LRF (laser range finder). Generally, existing localization technologies for the mobile robots used encoder, gyro, geomagnetic sensor, and so on. However, those local sensors for localization have errors accumulated by system characters of sensor, or external environment. To solve the cumulative error, the sensor fusion using several local sensors and probabilistic methods has been studied. However, these methods may cause high computational cost, and is unsuitable for the mobile robot which needs real-time control. Therefore, this paper proposes localization error correction method with low computational complexity on mobile robot using LRF. For experiment, we designed and made the mobile robot which has two encoders, a gyro and a LRF. We measured localization errors while robot moves on fixed path, and compared results of the proposed method with results of the sensor fusion using only gyro and encoder. In the experimental result, we verified that the proposed method is able to reduce localization error effectively.

抄録全体を表示

In this paper, we present a visual navigation method in an indoor GPS environment using D* Lite and fuzzy motion planning for humanoid robots. First, we estimate the location of the robot using an inner GPS(iGPS). A global path is searched considering the direction to the goal based on D* Lite planning method. A fuzzy motion planner finds feasible locomotion primitives to track the global path. It selects an optimal primitive depending on a distance between the path and the humanoid robot and an angle difference between a tangent angle of the path and a pointing angle of the robot. Fuzzy motion planner also decides that selected primitive can be applied to the walking motion of humanoid. If it is feasible, a humanoid robot executes locomotion primitives in order to navigate to the goal along the path. We verify our approach through the navigation experiments by using a developed small humanoid robot. The experimental results demonstrate that the humanoid robot can navigate fast and stably to the goal.

抄録全体を表示

This paper presents a design framework based on a GPS(Global position system) and Bluetooth communication for leader following and migration in decentralized control system. Very few attempts have been made at the study of the leaderfollowing system with combination of a GPS and Bluetooth communication. As well, their leader following systems are not decentralized by the use of a central computer. In this study, the leader robot migrates to the destination by the GPS receiver and magnetic compass sensor. The follower robot receives the position information of the leader robot through a Bluetooth communication between the leader and follower robots. And the current position information of the follower robot is received from its own GPS.

抄録全体を表示

In this research, a crab robot is implemented in H/W based on Jansen mechanism, and its kinematics is analyzed. The proposed biologically inspired robot is applicable to solve many engineered problems, since it is able to walk with maintaining its body still. The biologically inspired robot is able to achieve the mission appointed by a programmer connecting with outside, based on Blue-tooth communication module. Therefore, the purpose of this research is the implementation of the biologically inspired robot that can operate most adaptively in sand and marsh. For the purpose, the twelve-bar linkage of a crab robot is analyzed kinematically.

抄録全体を表示

Java Card technology provides interoperability, platform independence and high level security, java language has inherited problems that are slow execution speed and data loss particularly when power is gone while recording.So Java Card System first backups the previous data aboutJava objects into the transaction mechanism, then writes new data into Flash, during the downloading and execution of Java Cards application called an applet.It is caused by too many Flash writing operations.It makes Java Card much slower.In this paper, we have proposed a method to reduce the number of writing that new transactionmechanism in RAM. With our approach, the number of Flash writing and the downloading speed are reduced by 30% and 50% separately.

抄録全体を表示

Research on extracting technology of structure information from comparison sentences are to extract comparison target, features, and relations from given sentences and it is a very important technology in comparing search. There have been studies in English speaking countries, but this is the first study on Korean documents. In this study, the comparison targets and relations were focused among the structure information of comparison sentences to find the superiority and inferiority determination rules. Then, the comparison keywords for superiority and inferiority as well as the predicate were used to extract superior targets among the comparison targets. Termsb boda (than)bihae (compared with)hweolssin (far), and deo more were used for experiment and 93.9% of accuracy was resulted.

抄録全体を表示

According to increasing of cancer-related researches, various cancer research subjects made a lot of candidate cancer relatedgene sets. However, currently there are few independent gene set test tools which provide independent gene set test. We provide a convenient and intuitivetool(WittyCG) for independent cancer gene set test.WittyCG(Witness to your Cancer Gene sets) supports independent gene set tests with various tissues and various experiments types. In addition WittyCGillustrates the result of gene set test graphically with statistical information.

抄録全体を表示

Microarray data contain the expression levels of thousands of genes in multiple samples. Biologists try to find interesting patterns in gene expressions which show significant characteristics such as meaningful contrast in compared groups. It is one of important issues to identify the contrasting gene or sample clusters with respect to the compared group. This kind of analysis provides information about the biological markers in disease development, diagnosis, recurrence, or progression. Due to the high dimensionality of microarray data, it is not easy to find out which gene or sample clusters show contrasting behavior. This paper proposes a method to mine all contrasting clusters in microarray data. It makes use of fuzzy techniques to determine contrast of expression data, and incorporates an association rule mining technique which takes into account the curse of dimensionality.

抄録全体を表示

Based on the analysis of the interaction between a manipulator's hand and a working object, a model representing the constrained dynamics of the robot is first discussed. The constraint forces are expressed by an algebraic function of states, input generalized forces, and constraint condition, and then direct position / force controller without force sensor is proposed based on the algebraic relation. To give the grinding system the ability to adapt to any object shape being changed by the grinding, we added estimating function of the constraint condition in real time for the adaptive position / force control, which is indispensable for our method instead of not using force sensor. Evaluations through continuous shape-grinding experiment by fitting the changing constraint surface with spline functions, indicates that reliable position / force control and shape-grinding work can be achieved by this proposed controller.

抄録全体を表示

Hindrances interfering realistic and practical utilization of hyper-redundant manipulator is thought to be the facts that the higher redundant degrees make the weight of the structure heavier, resulting in a conflict between the required high-redundancy for dexterous manipulation and heavy weight stemming from the high redundancy. To overcome this contradicting realities of hyper-redundant manipulator, we discuss a realizability whether the contacting and bracing motion of intermediate links with environment may simultaneously reduce energy-consumption and raise hand's trajectory tracking accuracy, inspired by human's handwriting motion with the elbow or wrist contacting to a table. We first propose a dynamical model of hyper-redundant manipulator whose plural intermediate links are being braced with environment, second how the contacting can save energy and raise trajectory tracking accuracy. The simulation result shows that lessenergy consumption and high trajectory tracking accuracy are achieved by conventional PD controller, compared with noncontacting manipulator condition.

抄録全体を表示

In this paper, we focus on how to control the robot's end-effector to track an object, meanwhile, to approach it with a desired posture for grasping, which we named as approaching Visual Servoing (AVS). AVS with binocular cameras needs inheritably eye-vergence motion to keep cameras' sight against the target since approaching motion makes the cameras sight narrow or losing the object. We propose a hand & eye-vergence dual control system to perform AVS. The experiment of full 6-DoF Approaching Visual Servoing experiments to a moving object by a 7-link manipulator installed with a binocular camera system, confirmed the ability of Hand & Eye-vergence control system.

抄録全体を表示

This paper is concerned with a new concept of avoidance manipulability inspired from manipulability. The manipulability represents the ability to generate velocity at the tip of each link without any designated hand task. The avoidance manipulability represents the shape-changeability (avoidance ability) of each intermediate link when a prior hand task is given. The intermediate links represents all comprising links of robot except top link with end-effector. The avoidance matrices, 1Mi (i=1,...,n-1) corresponding to i-th intermediate links, are used for analyzing avoidance manipulability, resulting in that rank(1Mi) declares the shape-changeable space expansion and singular values of 1Mi indicates the avoidance ability of i-th link. In this research, what assumption can guarantee mathematically the configurations with maximum rank(1Mi) is our main concern for maximizing shape-changeability to prepare effectively dynamic change of environment or sudden appearance of obstacles. Then we proved that our "non-singular Configuration Assumption" can guarantee the maximum rank of 1Mi through detailed decomposition analyses of 1Mi.

抄録全体を表示

This paper is concerned with a concept of reconfiguration manipulability inspired from manipulability. The reconfiguration manipulability represents a shape-changeability of each intermediate link when a prior end-effector task is given. Taking reconfiguration manipulability into consideration, we analyze the redundant robots in view of its structure and shape by simulations.

抄録全体を表示

In this study, the process of induction of mutual cooperation behavior in workers in a remote work environment (Telework) was examined experimentally. In the experiment, the gambling task developed by Payen was used, and the subjects were not given the honor information of their partners. In addition, the subjects and their partners were kept in different rooms in order to avoid the influence of conversation. The subjects were only informed about which button their partners pushed. The number of trials was 100. The results show that cooperation behavior was induced in about 20 trials. Even if honor information of the partner was not known by a subject, cooperation behavior was induced by the information about which button the partner pushed. In the control experiments, the subjects were matched with a computer. Results of the control experiments suggest that cooperation behavior was not induced when the human subject was matched with a computer. This suggests that an exchange of information that a human being cannot express by a word (non-verbal information) was necessary for induction of cooperation behavior.

抄録全体を表示

Brain-Computer Interfaces (BCIs) control a computer or a machine based on the information of the signal of the human brain, and the P300 speller is one of the BCI communication tools. The P300 speller discriminates a character after averaging 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. This paper proposes Reliability-Based Selective-Repeat ARQ (RB-SR-ARQ), which selectively requests a user to re-send the data based on the reliability of each data. The results show that the time required for thought discrimination can be reduced while the accuracy remains at a high level.

抄録全体を表示

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 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.

抄録全体を表示

It has been reported in many papers that the number of eye movements increases and the mean fixation durations on individual objects decrease when the road situation becomes more complex. In this paper eye-movement analysis has been used for understanding mechanism of visual search of drivers. For the aim, two experiments in which the subjects drove the driving simulators have been conducted. We discuss the influences of the peripheral vision for the eye movements characterized by averaged frequency of saccadic eye movements and mean fixation duration.

抄録全体を表示

In this paper, it is examined to clarify the relation between feature of eye movements and driving behavior in conventional and unconventional situation by the experiments using the driving simulator and the eye tracking system. The experimental system proposed in this paper is useful for analyzing human error induced from the different traffic regulation between Japan and USA scenarios. It is found from the time histories of eye movements by synchronizing vehicle trajectories when the overshooting in the right turn and undershooting in the left turn that resulted in head-on crash, the duration time and number of eye fixation with overshooting or undershooting became larger than that without overshooting and undershooting.

抄録全体を表示

We numerically examine genetic algorithms (GAs) to determine the optimal set of the weights for three essential interactions in BOIDS under `zero-collision' and `no-breaking-up' constraints. As a fitness function to be maximized by the GA, we choose the so-called the `Gamma-value' of anisotropy which is empirically observed in typical flocks of starling. We confirm that the GA successfully finds the solution having a large gamma-value leading-up to a strong anisotropy. This result tells us that the BOIDS modelling of the flocking with the assistance of the GA might simulate a realistic flocking automatically even in our personal computers.

抄録全体を表示

The objective of this paper is to solve the dynamic plane covering problem by the movement of multiple robots, such as sprinkling water to a large field by several vehicles or aircrafts. In the problem, all of the points in the field should be covered with an equal density. One of the ways to solve the problem is the swarm leading control method, in which one robot, called a leader, moves along a path in the field, and all the other robots, called followers, move around the leader with a fixed distance. However, because the leader and the followers have asymmetrical relations, the stability condition of the swarm leading control method is unknown. In this paper, introducing a Manhattan norm for distance measurement, the stability condition can be shown and proved. The proposed control method is verified with computational simulations.

抄録全体を表示

Cloud computer technologies are advanced better, and web service like IaaS is used as businesses for its easiness and flexibility. But people concern about security on the cloud service and still feel resident to storing enterprise data in external server. So people interest in private cloud environment. In this paper, we evaluate private cloud system including Eucalyptus, an open-source software implementation of cloud computing system like Amazon EC2. Evaluations are about basic performance, management and recovery, web application performance for this system. Finally, we conclude if limited to small-scale, enterprise private infrastructure for short time use, it can be a good private cloud system

抄録全体を表示

An adaptive nutrient control method for a plant factory is proposed. The method is based on a Reinforcement Learning modified for a target in which one state never comes back during a single episode and a reward is given after a very long delay. In applications such as plant growth control, one episode takes a very long time period, and a rapid convergence to a prospective control solution is essential, whilst an extensive exploration is needed since there is usually no precise model available. A method like Reinforcement Learning is useful for a problem having no reference model. But a necessity of exploration does not match the need for rapid convergence, and a new balancing method is needed. In this research, an average reward distribution method is proposed, which is similar to the Profit Sharing method but affects more extensively to find much prospective early solutions, whilst guaranteeing to converge into a rational solution in a long run. An experiment is conducted in a simple plant factory system, which shows that at least standard Reinforcement Leaning is insufficient for this type of problem. Computer simulations show that the method has good effects comparing to a standard RL and a profit sharing method.

抄録全体を表示

This paper proposes a framework of a distributed kinematics resolution model which can be applied to various types of multi link robot structures. In the framework, hardware modules which consists of joints and links are defined as joint components of robotic systems, then kinematics resolution models are composed of a set of the local kinematics modules (LKMs) which are software module to compute the localized direct and inverse kinematics corresponding to each joint component. Kinematics computation throughout the robotic systems is resolved by the localized kinematics calculation of LKMs for an end effecter to a target position and an information exchange among LKMs. The proposed framework has been applied to case studies of the inverse kinematics problems of 7 degree of freedom multi link robot manipulator. Case studies have confirmed the proposed model to be able to provide a dexterous framework of kinematics resolution systems for modular type robotic systems and various types of robotic systems.

抄録全体を表示

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 multiple objective linear functions and linear constraints involving random variable coefficients, P-Pareto optimality concept is defined in cumulative distribution function space. After each decision maker specifies his/her own decision power and reference probabilistic values, the corresponding candidate of a satisfactory solution is obtained from among P-Pareto optimal solution set on the basis of linear programming. Interactive processes are demonstrated by means of an illustrative numerical example.

抄録全体を表示

In this paper we consider a multiobjective two-level integer programming problem in which there is not cordination between the decision maker at the upper level and the decision maker at the lower level. The decision maker at the upper level has an objective function and the decision maker at the lower level has multiple objective functions. The decision maker at the upper level must take account of multiple rational responses of the decision maker at the lower level in the problem. We examine two kinds of situations based on anticipation of the decision maker at the upper level; an optimistic anticipation and a pessimistic anticipation. Mathematical programming problems for obtaining the Stackelberg solutions based on two kinds of anticipation are formulated and three kinds of computational methods for obtaining the Stackelberg solutions are proposed. We carry out numerical experiments in order to demonstrate feasibility and effectiveness of the proposed methods.

抄録全体を表示

In this paper, several different multi-agent based simulation systems are constructed using some types of learning mechanisms for the market entry games which are played by n players. This paper extracts some learning and behavioral rules of human in the market entry games by comparing the results of laboratory experiments using human subjects and the simulation experiments.

抄録全体を表示

Heat load prediction in district heating and cooling (DHC) systems is one of the key technologies for economical and safe operations of DHC systems. The heat load prediction method using a three-layered neural network proposed by the authors has been used in an actual DHC plant on a trial basis. However, there exists a drawback that its prediction becomes less accurate in periods when the heat load is nonstationary. In this research, we attempt to improve the heat load prediction method through the introduction of a recurrent neural network using particle swarm optimization (PSO) for adapting the dynamical variation of heat load together with a new kind of input data in consideration of the characteristics of heat load data.

抄録全体を表示

This research considers Stackelberg solutions for two-level linear programming problems under fuzzy random environments. To deal with the formulated fuzzy random two-level linear programming problem, alpha-level sets of fuzzy random variables are introduced and an alpha-stochastic two-level linear programming problem is defined for guaranteeing the degree of realization of the problem. Taking into account vagueness of judgments of decision makers, fuzzy goals are introduced and the alpha-stochastic two-level linear programming problem is transformed into the problem to maximize the satisfaction degree for each fuzzy goal. Through expectation optimization in stochastic programming, the transformed stochastic two-level programming problem can be reduced to a deterministic two-level programming problem.

抄録全体を表示

In this paper, we propose the intelligent surveillance system using network cameras. To do this, we use two Internet Protocol (IP) cameras, which are installed at the inside and the outside of the restricted area. First of all, the external camera is used to monitor a potential intruder entering the restricted area in real time. To detect the potential intruder from sequential images obtained by external IP camera, we propose the robust algorithm combining the background modeling, the hybrid silhouette, the convex summation, the morphology and the smoothing. Once the intruder is detected by the external camera, the internal camera with pan-tilt mechanism is then activated and starts tracking the intruder and monitoring the prescribed valuables registered with some feature points in the database. To track the intruder, we propose the tracking algorithm using the background elimination and the histogram. And then, we completes the proposed surveillance system by using the Scale Invariant Feature Transform (SIFT) algorithm in order to monitor whether the intruder steals the prescribed valuables or not. Finally, we have shown the effectiveness and the applicability of the proposed method through experiments.

抄録全体を表示

In this study we augmented a virtual object to an image of a flexible surface such as a paper, which is acquired from a web camera. To get more presence feeling, we consider realistic illumination on augmenting. To get the geometric relation between the camera and the flexible surface, we use markers that are printed on the surface. Using marker information in prior, three dimensional coordinates of the surface can be calculated. After the marker is removed from the input image, we attach a two dimensional texture and shadow to the flexible surface by considering realistic illumination.

抄録全体を表示

This paper presents to study on simplifying the face image for encoding the ID picture to the standard 2D barcode on the business card. In the modern society, the business card is used to notify others of self-identity. Because of that reason, it has usually been exchanged when people meet others for the first time. However, they often forget the owner of the business card because of getting the cards more than more. To solve the problem, many applications were developed with smart-phone or internet. Those applications provide in various design and information. However, they didn't provide the convenience and practicality as existing(=paper) business card. Therefore, we study a method about the change of ID picture on a business card through simplifying ID picture to 2D bar code. In the experiments, we compared the original face image with created face image using cubic spline interpolation defined by feature points. We reduce the data as small as possible, still it can be expressed by 2D barcode, and confirmed that it can effectively express feature.

抄録全体を表示

This thesis proposes surveillance system to prevent the crops stolen. In most of the security system, a supervisor has to observe video at first hand. However, to overcome above limited ability, the proposed system realizes an intelligent surveillance method which can perceive confronted situation by detecting a moving object in wide area via vision technology. Although vision technology provides images, to identify the detected object has limitation due to low resolution. Therefore, the system is built together with a smart card.

抄録全体を表示

In case of the scanning document images, they are mostly the low-resolution images and the skew images. Many document images can be applied to a variety of pattern recognition area. The document images need image preprocessing including the process of rotation. In this study, we used of three rotation processing methods: one-pass method, two-pass method, and three-pass method, for the skew coordinate transforming of the document images. We compared the result images of three rotation methods and analyzed the time complexities of three rotation methods. In this paper, we took better results in the three-pass rotation method for the document images under the angle condition between one and fifteen. And the three-pass rotation method had better performance than the one-pass rotation method.

抄録全体を表示

Ship mates should be aware of images, numerical values, texts & audio-based information of radar, AIS, NAVTEX, and VHF, etc. for safe navigation; however, it is very complicated and difficult for them to acquire such information and use it as data for decision-making for safe navigation while keeping watch for navigation. So, a system to understand, unite and provide multimedia marine information for mates in voice is necessary. This study tries to suggest knowledge representation models necessary for the construction of the system to understand, unite, and provide the multimedia marine information for mates.

抄録全体を表示