本研究の成果は，大きく二つである．ひとつは色彩感情，五感の印象，および学習感情に関わるオントロジーの構築である．もうひとつは共感演出の検証である本研究で構築されたオントロジーでは，心理的属性に関わるいわば心の内容を扱ったものである．限られた文脈が限定されるものの，心理的反応のオントロジー記述としては非常に新しい試みである．このオントロジーを援用して共感を演出する際に，ドメインモデルの規約を与える概念を提供することができると考えられる．

本研究では、人間の感じる「調和感」の生成メカニズムを検討した。特に、色彩を中間言語として、様々な感性情報（視覚、聴覚、嗅覚等の感覚）を結びつけるための基礎的な研究とその実環境への応用に関する研究を行った。その成果は以下の通りである。(1)「調和感」生成に関する基礎的な知見を得たこと、（2）調和空間の構築に関する比較的普遍性の高い知見を得たこと、(3)「調和感」生成に関する脳科学な観点からの知見を得たこと、(4)「調和感」の統合的管理のためのオントロジーを構築したこと。なお、これらの成果に基づいた「調和感」の統合的管理システムの構築が今後の発展的課題である。

本研究の目的は,対象についての印象評定に関する知識を抽出し構造的に管理すると共に,それを利用した印象評価手法の開発である.知識を構造的に管理し利用するためには,オントロジー工学の方法論による知識の体系化が有用である.本年度は印象の評定に関する知識を記述するためのオントロジー構築を行い,次の成果を得た.
(1)色彩感情を測定する際に用いる評定尺度が,記述的次元の評定尺度と評価的次元の評定尺度の二つに分類されることを実験的に検証した.因子分析の結果からそれら二つの次元の評定尺度が区別されることを確認すると共に,快-不快,覚醒-睡眠,支配-服従という感情の次元を表す評定尺度は記述的次元と共通の性質を持ちながらも区別されることが示された.その上で,記述的次元と評価的次元の評定尺度においてどれほど,快-不快という評価の意味が共有されているかについて実験的に検証すると共にオントロジーとしての表現を試みた.
(2)Self-Assessment Manikin尺度を用いて単色に対する印象評価の実験を行い,構造方程式モデリングによって明度,彩度,色相という色彩の三属性と,快-不快,覚醒-睡眠,支配-服従という感情の三次元の関係を明らかにした.その上で,得られたモデル式を記述するための概念を整理した.
(3)色彩感情を表す評定値のような主観的データの記述・共有を目指し,既存の上位オントロジーに基づいて主観的な質・量に関する概念の定義を行った.また,色彩研究でよく用いられる色の表現方法である色空間モデルに関する概念と共に,主観的データの記述である統計モデルに関するオントロジー記述を行った.
以上により,印象の評定結果を記述するための概念的基盤として,オントロジーを構築することができた.このオントロジーを参照することによって,人間が抱いている印象(を表す知識)を概念レベルで管理・共有することが可能であると考えられる.

<p>In recent years, composers are required to have a high level of understanding of musical equipment due to the diversification of equipment, and researches on musical instruments and timbre understanding support systems have been conducted. Some of the previous studies use the overtones of the instruments as features and support them with illustrations based on principal component analysis, but further investigation of the correspondence with the overtone parameters is needed. In this study, the similarity of the overtone components was calculated using the dynamic time warping method and illustrated by clustering and principal component analysis, and the correspondence between the parameters was discussed in detail. As a result, it was suggested that there may be a correspondence between the visualized figure and the fundamental frequency components in the middle and high frequency range.</p>

<p>In this study, we examine influence of difference in estimation method of point of gaze on evaluation of transparency during watching see-through vision. There are the two methods of estimating point of gaze, which are the method that uses the midpoint of the line segment that shows the shortest distance between the two lines of sight in three-dimensional space or the method of estimating it from intersection of two lines of sight on the two-dimensional plane. We compared these methods. As the result, there are significant differences in estimation error of convergence distance and there is relatively small variance in the three-dimensional point-of-gaze estimation methods. This result suggests that the three-dimensional point-of-gaze estimation method is more effective in evaluating transparency.</p>

<p>When using a driving simulator, there is a problem that if the driving environment is not realistic enough, there will be a deviation from actual driving. On the other hand, it is said that using a head-mounted display (HMD), which has become widespread in recent years, improves the sense of reality in a VR environment.</p><p>In this study, for the purpose of creating an index to evaluate the sense of presence, we simulated driving under two conditions of HMD and flat display, and classified the factors from subjective impression evaluation.</p>

<p>In this study, we conducted an experiment using the SD method questionnaire to find out how the impression of the person who touched the line is affected by the difference in the three-dimensional shape of the line formed by the laminate shaping method. We conducted a factor analysis on the questionnaire results and compared the impressions of each line. As a result of the factor analysis, five factors were extracted, the tendency of the line impression was found from the result of the questionnaire, and a significant difference was found in the impression difference of some lines. In addition, the relationship between the adjective pairs used in the SD method and the five senses was compared, and it was suggested that the lines used this experiment give the impression of being highly related to vision.</p>

The olfactory brain that controls olfaction overlaps with the emotional brain that controls emotions, and is thought to have contributed greatly to the development of the emotional brain. Therefore, the sense of smell is considered to be more connected to emotions than other senses, and attention has been focused on manipulating the psychological state to a desirable state by using scent. However, it is thought that there is a gender difference in aroma. According to Oloveira et al., Women have a large number of neurons and neurons in the olfactory bulb, which play the role of receiving olfactory information, and women are superior in the smell separation test (1). Other studies have found that repeated exposure to odors increased olfactory sensitivity only in postpubertal women. From the previous case, it is considered that there is a gender difference in the intensity and type of emotion that occurs even when the same scent is presented. Therefore, in this study, we aimed to clarify the gender difference of psychological changes caused by scents.

This study aims to examine the relationship between visual fatigue and prefrontal cortex (PFC) activity during a backlit text reading task. We conducted a 10 min luminescent sentence reading task in a dark room. Time-course of oxy-Hb in the PFC was measured using a 22-channel wearable near-infrared spectroscope (NIRS). Subjective fatigue was assessed using the "Jikaku-sho shirabe" questionnaire. The critical flicker-fusion frequency (CFF) and LF/HF were assessed as objective index of fatigue and autonomic nervous system. 192 trials consisted of twelve healthy young volunteers and twelve LED treatments were analyzed. In the analysis, a linear non-gaussian acyclic (LiNGAM) model for causal discovery was introduced in addition to the correlation analysis. Our results suggested that the development of visual fatigue and central fatigue is related to the deactivation conditions in the medial PFC and Left PFC.

<p>Mobility scooters are three or four-wheeled electric-motor-driven wheelchairs with a handlebar for steering. In the use environment of a mobility scooter, there may be many obstacles that need to be avoided, such as pedestrians on a narrow sidewalk or cars on the edge of the road. These obstacles not only interfere with smooth driving but also induce mental strain on the user such as fear and anxiety, which may cause a decrease in comfort and an accident. The purpose of this study is to estimate physical obstacles in the road environment when driving a mobility scooter from operation information and physiological information of the user obtained during driving. We measured skin conductance, manipulation amount of accelerator, steering torque, and acceleration value in the horizontal plane when driving on a few courses with physical obstacles. We tried to build a machine learning model that estimates the type of physical obstacles in the course based on these time-series data by using a random forest classifier.</p>

<p>This study aims to examine the benefit of visual fatigue evaluation based on prefrontal cortex (PFC) measurement by near infrared spectroscopy (NIRS), and to evaluate fatigue on various colored LED backlit text. In our method, eight adult participants read a LED backlit text 10 minutes in a dark room. Before and after task, subjective fatigue questionnaire "Jikaku-sho shirabe", subjective visual fatigue questionnaire (SVF), and the critical flicker-fusion frequency (CFF) were measured. The time-course of oxy-Hb in the PFC was measured using a 22-channel NIRS. Heart rate variability (HRV) was measured to access an autonomic nervous system balance. Six types of white LEDs were used for backlighting: general white as a control (W), high color-rendering white (WH), bluish white (WB), reddish white (WR), greenish white (WG), and yellowish white (WY). In our results, correlation analysis suggested that the development of central fatigue is related to deactivated condition in the mPFC, and the development of subjective visual fatigue is related to deactivated condition in the LPFC. In comparison between LED treatments, in WH condition, high subjective central fatigue, CFF decay and LF/HF were revealed, and the PFC inactivated. In WG condition, low subjective central fatigue and CFF decay revealed. Our findings suggested that participants get tired on high color-rendering white LED backlit text, and were less tired in greenish white LED backlit text. Also, NIRS provides benefits in the evaluation of visual fatigue.</p>

<p>Obesity has become a serious problem worldwide in the modern society with the change of food culture. Obesity would be a cause of many diseases such as hypertension, dyslipidemia and diabetes. Considering obesity, we examined satiety closely related to energy intake. In this study, we focused on the number of chews and chewing frequency, which are factors of satiety, and analyzed the influence on it. In the experiment, we compared satiety, hunger and masticatory muscle activity among the conditions when participants ate the same amount of food under the condition that the number of chews and the chewing frequency were changed. As a result, the datas of the increase rate of satiety and the masticatory muscle activity showed that participants of low weight and normal weight who consumed foods by chewing many times, and participants of obesity who consumed foods with large masticatory muscle activity for each chew could feel fullness earlier. Using this result, people can reduce the energy intake naturally by feeling fullness early in the diet, and it is expected to improve the eating habits of obese people.</p>

<p>In recent years, new indicators focusing on cognitive activities have been developed for the purpose of improving the accuracy of lie detection tests. It has been pointed out that eye movement is particularly effective. In addition, there have been many reports of lie detection based on gaze movements and blinks. But these indicators are greatly influenced by human intentional control. In this study, we focus on the pupil diameter, which is difficult for humans to control intentionally. we verify whether it can be used a new index of lie detection method. In this experiment, we analyzed changes in pupil diameter when presenting a critical question and a non-critical question in a CIT issue. As a result, when the pupil diameter was dilated at the time of question presentation, the pupil diameter at the time of presentation of the critical question was smaller than that at the time of non-critical question presentation. This is because the cognitive load is suppressed by making the participants unaware of the contents when the critical question is presented. This result suggests that the pupil diameter can be used as a new index in the lie detection method.</p>

<p>When Fragrance influences the psychophysiological state of a person, and consequently the volume of fingertip plethysmography is affected by the aspiration of the aroma. Recently, studies have attempted to employ aroma in various scenarios, such as in smells for medical care, relaxation, environment improvement for productivity enhancement, and others. However, the type of fingertip plethysmograph behavior produced near the stimulation threshold that allows feeling the smell remains unclear, especially near the discrimination threshold, where fragrance intensity changes. In this study, the stimulation threshold at the time of aroma suction and the emotional change around the discrimination threshold are evaluated by analyzing both a questionnaire and fingertip plethysmograph measurement. As human biological systems produce chaotic behaviors such as heart beat rhythms and the plethysmograph, chaotic time series analysis was performed for fingertip plethysmograph measurements. Then, we evaluated psychophysiological responses during aroma suction, using a feature of the chaotic behavior in fingertip plethysmograph. In addition, we analyzed and compared with the conventional method LF / HF. As a result, we confirmed the fluctuation in the fingertip plethysmograph even when the research participants responded that it did not smell in the subjective evaluation. Further, we confirmed the fluctuating tendency of the fingertip plethysmographs when the Largest Lyapunov exponent increased as the aroma hen the maximum Lyapunov exponent increased as the aroma the fluctuation of the fingertip plethysmographs becomes large and corresponds to the uncomfortable values in from the subjective evaluation. In addition, the relaxing effect by smell was confirmed using LF / HF.</p>

<p>This study aims to examine the time-series effect of cognitive load on the cerebral blood flow. We used machine learning to extract features related to cognitive load in any brain region and time domain of oxy-Hb measured by NIRS. We conducted a 5min auditory stroop task. Ten healthy young volunteers participated in this study and performed five repeated experiments. The brain function in the prefrontal cortex was measured using 22ch wearable near-infrared spectroscopy. Cognitive load was assessed using questionnaire surveys NASA-TLX and accuracy. The results of analysis using neural network (NN) and layer-wise relevance propagation (LRP) suggest that around the inferior frontal gyrus (IFG) contain important features for classifying cognitive load, and oxy-Hb in IFG may increase with cognitive load. Our LRP method can extract common features to oxy-Hb within the same group, visualize the effect of cognitive load on individual data.</p>

<p>In this research, we conducted to generate Convolutional Neural Network (CNN) models which can distinguish between cancer cells and normal cells in the large intestine under 11 input image size (number of pixels) conditions (50×50 px, 100×100 px, 200×200 px, 300×300 px, 400×400 px, 500×500 px, 600×600 px, 700×700 px, 800×800 px, 900×900 px, 1000×1000 px) for the purpose of clarifying the effect of the input image size on the cancer cells identification in colorectal cancer pathology image. When we conducted to train models, we calculated model accuracy by 5-fold cross validation in each condition. As a result, the accuracy rate increased as input image size increased in 3 conditions (50×50 px, 100×100 px, 200×200 px). However, the accuracy rate of the model trained by 4 conditions (700×700 px, 800×800 px, 900×900 px, 1000×1000 px) were lower than that trained by other conditions. These results are thought to be caused by setting the kernel size to the same value in all conditions. Therefore, we concluded that it is necessary to determine an appropriate kernel size and stride for each input image size which can extracted features included in the input image when we conduct to generate a model for the purpose of distinguish between cancer cells and normal cells in the large intestine.</p>

<p>In recent years, the number of traffic accidents has decreased due to the development of driving support systems. Driving assistance system supports to the driver, however when driving assistance is not appropriate, the driver may be bothered by the assistance and distrusted of the system. In order to provide appropriate support, a method that driver behavior prediction has been proposed. In this research, we measured seat pressure distribution and driving information using a driving simulator, and predicted driving behavior using a convolutional neural network. In this research, we predicted five driving behaviors: "Go straight ahead", "Turn right", "Turn left", "right lane change", and "left lane change". As a result, it was possible to predict with an accuracy of 65.4% in inter participant evaluation and 96.4% in overall evaluation.</p>

<p>In this research, to develop a gait feedback training system optimized for individuals where trainees can efficiently train features that do not satisfy ideal walking using a deep learning, we examine classification and features visualization of gait using Convolutional Neural Network (CNN) and Grad-CAM. In the experiment, the thumb-floor distance of right foot was measured when young people walked normally and when they walked with a brace, to limit their movement. Further, these data were clustered to 3 clusters using k-shape method. And these data were learned and classified as input data and using these cluster as the label. As the result, the accuracy was 86.07%. In addition, the part where the feature in thumb-floor distance appears were visualized as heat map using Grad-CAM and it is confirmed that usefulness for gait training.</p>

<p>Mobility scooters are three or four-wheeled electric-motor-driven wheelchairs with a handlebar for steering. In the use environment of a mobility scooter, there may be many obstacles that need to be avoided, such as pedestrians on a narrow sidewalk or cars on the edge of the road. These obstacles not only interfere with smooth driving but also induce mental strain on the user such as fear and anxiety, which may cause a decrease in comfort and an accident. The purpose of this study is to estimate physical obstacles in the road environment when driving a mobility scooter from operation information and physiological information of the user obtained during driving. We measured skin conductance, manipulation amount of accelerator, steering torque, and acceleration when driving on a course with physical obstacles. We tried to build a model that estimates the type of physical obstacles in the course based on these time-series data by using a random forest, which is one of the machine learning methods.</p>

<p>Smoke separation is accelerating towards the Tokyo Olympics and Paralympics. However, there is no method for monitoring harmful chemical substances in tobacco smoke in real-time, so there are scientific lacks against non-smokers health whether the smoke separated space have no impact or not to their health. The main chemical substance in tobacco smoke is nicotine. We decided to build up a model first for nicotine. In this study, we use a newly developing gas analyzer which measures nicotine in real-time. From that nicotine measurement result, analyzes the nicotine behavior with FEM in a simulated space. Then make it learn the parameters of the nicotine behavior model, and using a different modeling and learning algorithm than general multi-agent reinforcement learning, we visualize the nicotine behavior in tobacco smoke through the new achieving model. First, we confirmed the availability of nicotine analysis with the gas analyzer.</p>

<p>The purpose of this study is to clarify feature for discriminating emotion while expressing fake smile. First, we experimented to record face expression in video. The flow of the experiment was following; we first displayed one of emotion-eliciting films, and participant expressed (fake) smile just after watching the film. At the end of the experiment, we let them assess what emotion they felt with two self-report measures. We prepared six emotion-eliciting films in order to record various face expression with various emotion. Then, we obtained face landmark positions from the face on the recorded video using a method proposed by Kazemi et al⁽¹⁾. We focused that eyes and mouth corners move well while smiling, and defined 15 feature values related to that landmarks. Furthermore, we grouped the feature values in emotion scores by clustering. There are significant differences in eye area between happiness and disgust, and in some mouth corners' feature values between happiness and fear as a result of comparison of clusters. we considered these differences resulted from disgust/fear, but negative emotion.</p>

<p>The caution making is used in various daily situations as a preventive measure of tumble accident which is one of the labor accidents. The purpose of this study is to propose a display that incorporates the Fraser-Wilcox illusion and has a higher attention effect. Fraser-Wilcox illusion is a type of visual illusion in which a part of a still image appears to be moving. We compare the attention effects of shapes that induce the Fraser-Wilcox illusion and those that do not. As an evaluation method of attention recall effect, the pre-cueing task which was an experimental technique to measure the characteristics of visual attention was used. The correlation between the amount of visual illusion and the target reaction time was examined. As a result, a strong negative correlation was observed. The larger the amount of visual illusion, the faster the detection of the target. These results suggest that the Fraser-Wilcox illusion has an alerting effect on visual illusion.</p>

<p>The purpose of this study is to estimate the condition of second-hand goods from images using machine learning methods. Assuming that the condition evaluation of second-hand goods has a relationship with the damage that exists in them, we proposed and verified a method to estimate the condition of second-hand goods based on information of the damage. In this study, we chose men's dress shoes for condition evaluation and chose four types of damage: wrinkles, worn out of heel, dirt on outsole, dirt on inside as damage considered to be related to condition evaluation. (1)We create a CNN model that determines the presence or absence of damage, and (2) create a classification tree model that evaluates the state of the product from information on the presence or absence of damage and (3)We evaluated the condition of second-hand goods by combining two created models. As a result of verification, it was shown that the proposed method could evaluate the condition of used goods.</p>

<p>Warning at the time of switching from automatic to manual operation of vehicle is required to assist the driver appropriately. In this research, we focus on the driver's immersion in typing with a personal computer during automatic driving. We examined changes in driving operation after switching to manual operation when the warning sound is presented under four conditions including two conditions of beep sound and two conditions of presentation voice. The beeper was able to react quickly but it was found to be dangerous for driving operations. In voice presentation, it was suggested that a warning simply presented could not assist appropriately but presenting hazard avoidance instruction enabled quick reaction and assisted in calm driving behavior.</p>

Currently, the visually impaired find it difficult to use mobility support facilities because they depend on ambient environment such as ambient sound and snow cover. Further, people who acquired visual impairment in the later stages of life also find it difficult to use these facilities. Therefore, in this study, we aimed to develop a walking guidance system using 3D sounds that will not be influenced by the surrounding environment and will be easy to use. Further, we verified the induction effect in the blindfold state and the ambient sound. This system was implemented with a head mount display that is capable of producing 3D sound. Under a visual restriction environment, blindfolded participants were asked to walk toward a real or 3D guiding sound with three types of ambient sounds, and then, we measured the walking locus. We found that the 3D sound tends to provide veering tendency opposite of the ambient sound. However, the guiding accuracy of the 3D sound is similar to that of real sound.

<p>筆者らはこれまで，蝶の色彩美に焦点を当て，特に美麗種の多いアゲハチョウ科の配色傾向を調査してきた．このための主な手法として，蝶画像を色の類似度で分類し，各クラスターの色彩分布を示すグラフから配色傾向を分析してきた(梯ら,2018; Kakehashi et al. 2018)．本研究ではこれらのデータを統計的に処理するためにガウス混合モデル（Gaussian Mixture Model: GMM）による分析を試みた．GMMによって得られた結果を予め定義した配色型と比較したところ，主に次の三つの傾向が得られた．①明度コントラスト（低明度の面積の方が支配的），②類似彩度（低彩度の面積が支配的），③類似色相（橙～黄緑の面積が支配的）．これらの結果はOu and Luo(2006)の実験的色彩調和論をはじめとした先行研究の結果のうち，equal hue, equal chroma, unequal lightnessと一致していた．</p>

筆者らはこれまで，主観的な代表色抽出によってアゲハチョウ科の色彩傾向を調査してきた（梯,笠松 2015）．本研究ではより客観的なデータを得るため，類似画像検索技術を応用したアゲハチョウ科の色彩分析を目的とした．具体的には，Histogram Intersection（swain 1991）を用いて画像間の類似度を算出し，1-類似度を画像間の距離として階層的クラスター分析の変数に用い，108枚の蝶画像を分類した．各クラスターの画像群から，CIELChのヒストグラムを作成し，各属性の分布特性を分析した．その結果，主にL*とC*abの分布は50以下に集中し，二峰性分布では各峰のピーク間の色差ΔL*は50～90，ΔC*abは20～50または60～90であった．さらに，habの分布は30～120°または90～180°に集中し，二峰性分布でのΔH *abは60～150°または120～180°であった．これらの特徴をまとめると，アゲハチョウ科の蝶の色彩傾向は主に(1)低明度色や明度差が大きい配色が多く，段階的に明度が変化する． (2)低彩度色が多く，一部で彩度差が大きい．(3)色相は赤～黄緑や黄～青緑が多く，これらの補色がわずかに分布する．今後はより詳細な特徴を捉えるため，カテゴリカルカラーや三次元空間での分布傾向などについてさらなる分析を行う．

In this paper, we evaluated the mental stress recorded against corner entry speed while riding a mobility scooter by measuring skin conductance. A mobility scooter is a four-wheel electric wheelchair with a steering handle. These scooters, known as “Senior Car” are generally sold to the elderly. Some of these scooters have a system that automatically decelerates to approximately 70% of the maximum speed depending on the steering angle at corners. The preset deceleration rate can cause discomfort to the driver. To provide appropriate assistance without any discomfort, evaluating the mental stress of the driver is necessary. Skin conductance (SC) is one method for measuring sweating. It can measure mental sweating related to sympathetic nerve activity. We modified a commercially available mobility scooter to be used as an experimental wheelchair such that it can be controlled using a PC. We measured SC in a task that involved turning right into a corner at a constant speed using an experimental wheelchair and tried to evaluate the mental stress against the corner entry speed. Resultantly, a positive correlation between the amplitude of SC and entry speed was obtained. In addition, we performed a questionnaire to evaluate the ride discomfort after the experimental task. The questionnaire results demonstrate that the evaluation value of the items “No Fear” and “Safety” decreased as the speed increases. This result was consistent with the measurement results of the SC amplitude. From these results, we conclude that fear against corner entry speed could be evaluated using SC.

In recent times, robots such as, service, communication, and cooperative robots, have been developed to assist humans in various tasks. Smooth communication is an important parameter to ensure appropriate assistance by robots. In communication, understanding the comprehension status of the partner leads to an easy-to-understand explanation. However, it is difficult to accurately understand the comprehension status of the partner. The purpose of this study is to clarify the relationship between thinking status and pupil size and blink rate. Word association tasks were conducted, wherein subjects associated a theme from five auditory words. Pupil size and blink rate were measured during the task using gaze measurement. The result showed that pupil size tended to increase during thinking and decrease with association. Decrease in blink rate distribution was observed after presentation of a word, following which a temporary peak occurred before association. However, no change in blink rate was observed after the association. Based on the results of the present study, it was revealed that thinking may be associated with pupil size and blink rate, and it can be suggested that the eye could be an index to evaluate thinking status.

Determination of the optimal amount of information is essential to avoid confusion. In this study, the “Stroop task” was used for information processing, and the cognitive load was gradually increased by sequential adding of information. We designed and conducted two tasks that originate in the “Stroop task;” these two tasks feature significant differences in cognitive load. We also measured brain activity using near-infrared spectroscopy (NIRS) under the assumption that such activity could be used as an index of cognitive load. Both tasks were associated with increased oxy-hemoglobin levels in the prefrontal area, and the task with a higher cognitive load was associated with a more substantial increase in oxy-hemoglobin; this indicates that oxy-hemoglobin levels may be used as an objective index for the evaluation of information-associated cognitive load. In addition, this study included two groups (Group A and B) of patients who undertook the tasks. Based on the timing and the number of mistakes, it was observed that subjects in group A demonstrated lesser information processing ability than those in group B. The findings of the present study thus indicate that oxy-hemoglobin variation may depend on the information processing ability of the subject.

It is important to mesure energy metabolism for human health. Current mainstream activity meters show large inaccuracies in the estimation of energy metabolism for small mothion. In this study, we examined the estimation method of energy metabolism during exercise with small motion by using electromyography and heart rate, both of which are correlated with energy metabolism. We chose wrist curl a type muscular strength training to represent movement with small motion, and considered the flexor carpi radials and biceps brachii as target muscles. Multiple regression analysis was used to estimate energy metabolism; this allowed us to estimate that there is an error of about 25% when estimating total energy metabolism. Although comparison was carried out at ten-second intervals, estimation was difficult because of the large variation of experiment values. This study suggests that total energy metabolism during exercise with small motion can be estimated by using myoelectric potentials and heart rate.

At nursing care sites, patients undergo an exercise regimen as rehabilitation and training for walking disorders. However, it is presumed that their motivation decreases over time because in a nursing home, walking exercises are performed in a closed space and are repetitions of monotonous movements. Therefore, it is necessary to develop a more varied and effective rehabilitation or training method. The purpose of this study is to develop a varied and effective training method wherein the trainee walks while viewing virtual reality (VR) images. Specifically, in the proposed method, the trainee views VR images of walking on a slope, while actually walking on a treadmill, and he/she controls the physical load. In this study, we conducted an experiment using a composed training system with the aim of analyzing the influence of watching VR imagery of walking on a slope on the braking and driving forces while walking on a horizontal treadmill. We improved the VR system to automatically adjust the treadmill speed from the position information of the trainee. Then, comparative experiments of the braking and driving forces were performed for cases of walking while watching images of level and slope grounds. As a result, the walking speed, braking force, and driving force tended to decrease when watching downward slope images compared to watching level ground images. Additionally, the walking speed, braking force, and driving force tended to increase when viewing upward slope images.

Students in many universities, technical colleges, and technical high schools are practicing their cutting skills using general-purpose machining tools. However, there may be insufficient guidance due to the lack of the number of instructors and a limitation on the number of general-purpose machining tools that they can use. In this research, we propose a system to support the learning of cutting skills using a virtual reality head-mounted display. Preparing students to use machining tools require many training sessions, especially for cutting using a general-purpose lathe. A user can experience three types of content: drawing, tools, and the general-purpose lathe. This learning system was provided to seventeen students, and we conducted a questionnaire on its ease of understanding, appearance, enjoyability, intuitiveness, and operability. Resultantly, enjoyability and intuitiveness were highly evaluated. The results suggest that our proposed system can be used as a learning method. To improve the operability of the VR system, we can make use of specialized controllers or additional force presentation devices.

<p>It is generally agreed that the risk of falling increases with age. Falls in elderly individuals decrease their daily life activity level due to physical damage and increase fear of falling again. It is proposed that many elderly falls occur when walking, and in most, the cause is stumbling. Therefore, the aim of this study was to analyze the factors of stumbling during normal walking to prevent falls. Then we determined the thumb-floor distance, the angle of lower limb joints and the ground reaction forces. Furthermore, we calculated the moments of the lower limb joints from the obtained results. And compared these values during normal walking between elderly group and young group. Consequently, a significant positive correlation was found between the maximum value of the longitudinal component of ground reaction force and the maximum value of thumb-floor distance at terminal swing. Furthermore, when the longitudinal component of ground reaction force was the maximum, the hip flection moment of the elderly group was significantly lower than young group. These results showed that the cause of the fall of elderly is that the driving force at terminal stance is weak due to weak hip flexion muscles and that the knee joint is not sufficiently extended at terminal swing.</p>

<p>In recent years, it is required to design a notification sound which is easy to hear even in the daily life sound. It is necessary for the notification sound to use sound of frequency below 2000 Hz that the elderly can hear easily. Therefore, in this research, we aimed to evaluate how the influence on the recognition and impression of the notification sound is caused by the difference in the timbre. In this experiment, the participants heard three kinds of timbre of sine wave, rectangular wave and sawtooth wave. The threshold of listening sound pressure level was measured, and the questionnaire to evaluate impression of notification sound by semantic difference method was conducted. As a result, we found that the sine wave is familiar and does not give much tension or discomfort. In addition, we found that the rectangular wave and the sawtooth wave are less familiar and give a sense of urgency and discomfort. It was suggested that by changing the timbre of the notification sound, it is possible to design a notification sound that the intention is more easily heard and accurately transmitted.</p>

<p>Demand for vital sign measurement systems have been increasing. Non-invasive systems are required that it can monitor health status over a prolonged period. Conventional devices can easily measure pulses. However, as the noise due to perspiration is superimposed on the waveform of the pulse, the peak of the pulse becomes unclear, so it is unsuitable for measurement for a long time. In this study, we devised a pulse measurement method that can deal with the above problem. We obtained the acceleration waveform from the carotid artery skin surface, using the acceleration sensor. These waveforms are pulse waves caused by the expansion and contraction of the carotid artery. By the measurement procedure that they devised, six of ten people were able to be measured the pulse. In addition, it became possible to measure the pulse by other four people straightening the neck and attached the sensor to the appropriate position carefully.</p>

<p>In recent years, various researches on abnormality detection using machine learning have been conducted. However, when an abnormality is detected, it is necessary for engineers to find the cause of the abnormality and deal with it. If it is possible to estimate the cause of the abnormality at the time of detection, engineers can prepare appropriate repair tools and replacement parts. Machine downtime can be shorten by this estimation. In this research, we proposed new method and conducted experiment to verify its usefulness. In the proposal method, we use convolution neural network (CNN) to detect an abnormality and estimation a cause of it. Operating sound and current value were used for input. In the experiment, we chose geared motors as a model of a rotating machinery, and we used proposed method to them. As a result, we got an average accuracy rate of 82.41% on detection of abnormality, and also got average accuracy 67.84% on cause estimation of it. From these results, it was shown that the proposed method is effective for detection of abnormality and cause estimation of it for rotating machinery.</p>

<p>In recent years, it has been said that the number of mastication and mealtime tends to decrease with the spread of fast foods and softening of foods. Decrease in the number of mastication causes many problems such as obesity, shrinkage of the jaw, loss of natural teeth. In this study, we examined a method to increase the number of mastication by changing the subjective texture of food by presenting mastication sound with bone conduction headphones when chewing food, based on a cross-modal. Moreover, we investigated what kind of mastication sound is effective for increasing the number of mastication from the viewpoint of texture. The result showed that there was a possibility of increasing the number of mastication. On the other hand, a certain result was not obtained with regard to the texture, because the texture which has the correlation with the number of mastication was different depending on the food. However, as the result of a questionnaire of the texture, the correlation between the texture of heavy and the number of mastication was found. Therefore, we confirmed the possibility of increasing the number of mastication by controlling the subjective heaviness of food.</p>

<p>In recent years, cognitive support research has been conducted to reducing the number of traffic accidents. There is a possibility of confusing the driver if the cognitive support is not appropriate. Therefore, there is a need for a method to support driver's perception of dangerous objects without hindering normal driving behavior. In this study, we measured driving behavior and cerebral blood flow when moving sound was used as cognitive support. By using the moving sound as the cognitive support, the brake operation at the time of avoiding the danger became moderate. From the results of cerebral blood flow, it was found that information can be transmitted with the same cognitive load as in normal operation. As a result, the possibility that the moving sound can appropriately assist perception of the surroundings of the driver has been shown.</p>

<p>Light emitting diodes (LEDs) have been widely used for the luminescent text display in automobiles. The purpose of the current study is to evaluate fatigue reduction characteristics in LEDs luminescent text display. For this purpose, we construct a model to estimate the participant's fatigue from physiological index. In the experiment, participants read Japanese sentences that were represented by backlight emitting for ten minutes. Brain activity were measured for the physiological index, and fatigue evaluation questionnaire was answered. The luminescent text display was implemented with LEDs of six kinds of white color. We constructed the deep learning model that inputs brain activity, and outputs "fatigue", "no fatigue" or "no change". The output was calculated from fatigue evaluation questionnaire. The model was able to estimate the fatigue with a validation accuracy of 86.4 %. Thus, the model showed the possibility that estimate fatigue based on the brain activity.</p>

Typically, it is difficult to transmit visual images in words. The transmission of the visual image includes various information. Therefore, efficient communication is necessary. In this research, we focused on the relationship between onomatopoeia and modalism, and conducted a survey to measure the distance. We asked for general group and design skill holding group. Transmission of visual image in communication, we collected 200 words of onomatopoeia that are frequently used. Then we carried out relevance research on sensory modality, using the MD (Modal Differential) method. Then, we conducted a hierarchical cluster analysis and the semantic structure analysis which is called as SS analysis. As a result, we drew the SS graph for each cluster, which shows the relevance and order-related structure. The results were compared between the general group and the design skill preservation group, and discussions were made on the differences depending on the attributes of the subjects.

<p>The increase in the aging population is one of the common problems in the world. Many elderly people experience difficulty in standing up from their seats owing to muscular depression. To improve the QOL (Quality of Life) and ADL (Activities of Daily Living) of elderly people, improvement of the traffic environment is necessary. We developed an assistive seat to help elderly people rise from their seats in a bus with fewer loads. We verified the performance of the assistive seat through EMG analysis. In conclusion, the developed assistive bus seat was effective in helping the user to stand up without difficulty.</p>

<p>It is difficult for users of electric wheelchair who suffer from quadriplegia or have compromised motor skills to manipulation of the joystick finely. Also, there is a risk causing a loss of control upon impacts with steps. In this research, we design a system that performs to assist fine control of the speed depending on the height of the step. This system uses a laser range finder to detect the step from remotely, and uses the valuation of human exposure to whole-body vibration in order to estimate discomfort that arose from the step height and step climbing speed. The potential of this system is also verified and manifested to reduce the impact caused by collisions with steps.</p>

<p>Currently, Japan is entering a super aging society, and as the population of elderly people increases, the number of care recipients are also on the rise. The burden on nursing care homes is increasing; specifically, the burden on caregivers in excretion assistance is huge. Therefore, the need to develop nursing care equipment to support such tasks is increasing. In this study, a preliminary investigation was conducted to determine the possibility of excretion time prediction using data from the diet of a care recipient. We requested and obtained meal data and excretion data from experiment collaborators. We attempted to design a system that predicts the excretion time by a process similar to machine learning using a neural network that corresponds to the two sets of data. We can roughly predict the excretion time. The result showed that it is possible to predict the excretion time from the meal record. In the future, we will include details of the food we eat and biological information to improve prediction accuracy.</p>

<p>Sign language is used for communicating with deaf persons. To facilitate the communication between the deaf and the hearing communities, a sign language translator is required. However, sign language varies from country to country. Furthermore, the number of words is increasing as society develops, and as a result, artificial translation has become challenging. Therefore, substantial research has been carried out to recognize and translate sign language by means of computing methods. In this paper, we propose a sign language recognition method that can adapt to video images captured by general-purpose cameras, and not by special input devices such as a depth sensor. According to this method, sign language regions are extracted from moving images by interframe difference image processing, and are recognized by using a convolution neural network (CNN). As a result of the one-person-leave-out cross validation experiment to recognize 30 words, the highest validation accuracy observed was 90.0%. Thus, the proposed method's potential for sign language recognition is manifested.</p>

<p>In this paper, we aimed at obtaining knowledge about the effect of adding a virtual glass frame and additional information to the Virtual Reality(VR) image on VR sickness. The virtual glass frame is an object model imitating a glass frame arranged in the VR space. The frame is arranged to cover the peripheral visual field when viewing a VR image using a head mounted display. We created VR contents such a spin room and carried out comparison experiments between conditions where the virtual glass frame was added to the VR image using a questionnaire and bio-information. It was found that addition of the virtual glass frame to the VR image reduces the occurrence of vection in the peripheral vision and that the evaluation value of VR sickness decreases. Moreover, adding horizontal information to the frame reduces the confusion of vestibular sensation; furthermore, it was found that the evaluation value of VR sickness further decreases.</p>

<p>Automation level 3 of an automobile is an automatic operation in the normal state;, however, the operation is switched to the driver when it is determined that operation of the system is difficult to continue. In previous study, there were cases where automatic operation cannot be smoothly switched from manual operation owing to motion other than driving, or a decrease in arousal level during automatic operation. It is considered that estimating the condition of the driver can support smooth switching. In this study, we estimated the state of the driver while automatic driving using seat surface pressure. Results indicate that the estimation was approximately 100% in the individual evaluation and 78.4% in the overall evaluation using machine learning. This suggests that it is possible to estimate the driver's state while automatic driving.</p>

<p>Recently, some measures to prevent traffic accidents have been introduced; for example, information presentation by a head up display (HUD), vibration at the handle, and sound signals. However, the index to verify these utilities has not been established yet. According to a preceding study, the relationship between brain activity and negative emotion was observed by the near-infrared spectroscopy (NIRS). Moreover, other studies show that mental stress is related to galvanic skin reaction (GSR), respiration (RSP), and blood volume pulse (BVP). In this paper, we considered danger recognition to be associated with negative emotion and mental stress situation, and examined whether the status of danger recognition can be observed by combining bio-information. We observed that danger recognition is correlated to brain activity and GSR. Additionally, the oxy-Hb and GSR were observed to change when recognizing danger. This result suggests that measuring brain activity and GSR can be used to establish whether danger is recognized or not, and the potential of evaluating presence of danger recognition by using bio-information</p>

<p>LEDs are widely used for backlighting of text indicators in vehicles. It is desirable that such text indicators have high legibility and low fatigue-inducing qualities. Suitable spectral distributions for text presentation in vehicles have not yet been clarified. This study examined legibility and fatigue when participants watched various text presentations. Participants read Japanese sentences that were presented by backlight emission, for 10 minutes in a dark room. The backlight indicators employed two color LEDs. The participants were asked to answer questionnaires to evaluate legibility and fatigue. Critical fusion frequency, brain activity and electrocardiogram were measured to evaluate fatigue objectively. Based on the questionnaire results, both colored backlight equal in legibility. Subjective evaluations showed that fatigue is more likely to be caused by high color rending white backlighting. On the other hand, objective evaluations (critical fusion frequency, brain activity and electrocardiogram) showed that fatigue is likely to be caused by high color rending white backlighting.</p>

<p>It is said that music is effective for controlling feelings. For example, we can feel relaxed to listen to slow music. This paper proposes a generation system which composes music by means of a genetic algorithm. The composition which is generated by the system consists of the melody part, the chord part, and the bass part. The system can specify the atmosphere of the generated composition by choosing adjective words; "lightness", "intensity", and "complexity". To reflect the effect of adjective points on atmosphere of generated composition, the system adjusts the parameters of an evaluation function. As a result of a composition generating experiment, similar melodies and rhythms were found in each composition. To evaluate the appropriateness of the function about reflecting the effect of adjective points on atmosphere of generated composition, we conducted a questionnaire investigation. As a result, satisfactory answers are obtained for a specific adjective word.</p>

<p>In recent years, the number of people who suffer from dementia has increased as the aging population increases. Since a standard method for the treatment of dementia has not been developed, there has been an attempt to combine the treatment of dementia with complementary and alternative medicine. Among complementary and alternative medicine methods, music therapy has attracted attention because of its low risk of side effects. Currently, group therapy is the mainstream of music therapy in Japan because of insufficient number of music therapists. As a result, care for individuals has not been sufficient. Therefore, it is necessary to provide pleasant music for each patient that can be automated. For this reason, a music selection system is required, and this system can be realized by measuring emotions using biological information such as brain activity. In this study, we analyzed brain activity while listening to classical music using near infrared spectroscopy. In our experiment, participants listened to four types of music with different characteristics. The results show that the oxy-Hb rises when participants feel uncomfortable, while the oxy-Hb decreases when participants dislike the music. This indicates that emotions and preferences while listening to music can be identified using brain activity.</p>

<p>It is argued that, in human communication, the impression of a message is determined by nonverbal information. In this paper, we evaluated the relationship between human sensibility and eye movement. We conducted two kinds of experiments to induce a change in the psychological state. We used pictures of beverages as visual stimuli. A single image was applied in one experiment type, and double images were applied in the other experiment type. We carried out a questionnaire survey regarding the interest level with respect to the displayed image and measured eye movement with a gaze measurement system. The results indicated that the gaze time related to the displayed images with higher interest was longer in the double images experiment type. The number of blinks increased or decreased according to perplexity. The number of saccades increased or decreased in response to the perplexity levels in the double images experiment, and the change of saccades related to the interest in the single image experiment. These results showed that the eye movements have a potential for evaluating the psychological state.</p>

In recent years, with the development of IoT technologies, biological information of users has been easily acquired and accumulated. It is required that user's cognitive or emotional state could be accurately identified from such biological information. Various researches have used machine learning or deep learning however, generalized methods to handle biological data have not been established. In this paper, we propose a more versatile method using convolution neural network of deep learning to extract features from individual biometric information which is multi-channels time-series data representing user&rsquo;s cognitive. We demonstrate a new method of extraction and distinguishment of features from cerebral blood flow, skin temperature, respiration and skin conductance measured during specific task conditions in the stroop experiment.

<p>We developed a new power assistance suit without the need for electricity. This suit is to support care worker's assistance behavior, e.g. lifting-up and holding motion. It is made of only rubber and nylon belts. This suit can assist the equipped person's arms and back with two kinds of equipped belts. One is for the waist, attached from the shoulder to the back of the knee via his/her back. The other is for the arms, attached from equipped person's forearm to the back of the knee via his/her back. To confirm the possibility of this suit, first, we interviewed medical doctors and physical therapists. Their comments were, it is important to adjust the length of each belt for an equipped person adequately, and it is effective for the assistance of motion in a half-sitting posture and in transfer motions between a bed and a wheel chair. Furthermore, we measured the variation of maximum output force and %MVC of the entire body muscles during lifting and holding motions using the suit and not. From the result, by using this suit, the maximum lifting weight of the equipped person was increased by over 8kg, and the arms and back of the equipped person were assisted effectively.</p>

<p>Recently, value creation from view point of service has been widespread. A service process is regarded as a kind of interaction between service providers and receivers. Sharing and managing such the knowledge on human Kansei are of growing significance to providing product's values focused on customers. However, any sufficiently-advanced technologies that realize such knowledge sharing and management have been established. One typical approach is to clarify and systematize concepts as an ontology which is a basis to describe knowledge. In the current study, we discuss a method to develop an ontology describing knowledge on human Kansei and introduce several existing ontologies. The ontology will serve as a useful reference to develop value recognition map and be of assistance to designers.</p>

<p>We developed an automatic damage diagnosis system of the gear tooth surface using a laser beam. Only attaching a laser sensor in a gear box, we can grasp the size and place of the damage on a gear tooth surface without using a rotary encoder. We confirmed the limit of this method against the velocity fluctuation. Furthermore, to utilize this system in the practical field, we developed a new system which was small, light weight, and low cost. The accuracy between new and conventional system was compared, and we confirmed the results were comparable level.</p>

<p>Recently, improvement of design capability is one of important issues in industry However, it's really difficult to share knowledge of design possessed by experts with novices, because the knowledge exists depending on their skill and degree of prociency. Thus the current study aims to excavate the experts' Knowledge. We then described design objects related to action to design electronic components.</p>

<p>In recent years, various driving support system has been developed as traffic accident prevention. However, they are unless transmitted to the intention that led to providing support, there is a possibility that the rather confusing the driver. Therefore, when the adoption of the driving assistance system, consider plenty the human visual attention to contribute significantly to cognition, without disturbing with the normal driving behavior, it is required a method for supporting recognition of the surrounding dangerous object. In this paper, we investigated the reaction time when subjected to experiments using dual task, has been presented a variety of auditory stimuli as lead cue stimuli. As a result, when using auditory stimulation with sound image movement, it tended to the reaction time of the central issue and the peripheral issues becomes earlier were observed. The result showed that the possibility of inducing an attention without considering the driver by using sound image movement as a visual assistance.</p>

<p>In recent years, the number of dementia elderly person has been increasing. Various cognitive rehabilitation is underwent in nursing homes. However, it is difficult for them living far away from their home to undergo cognitive rehabilitation. Therefore, if they undergo cognitive rehabilitation in remote locations, many of them can improve dementia. In this research, we examined cognitive rehabilitation by cooperative between remote locations. When they undergo cognitive rehabilitation in remote locations, improvement of their presence is needed and it is assumed that cooperative operation can improve their presence. We chose a maze done by two persons cooperatively as concrete cooperative operation. In our experiment, firstly we compared brain activity between doing the maze and calculating because calculation is used as cognitive rehabilitation. Secondly we examined the possible of improvement of their presence during doing the maze. The result shows that the maze and calculation caused similar brain activity and cooperative operation improve presence. It reveals effectiveness of cognitive rehabilitation by cooperative between remote locations.</p>

<p>Various walking assistance devices which promote rehabilitation and exercise for elderly have been developed. However, these devices can't support walking for exercise, because many elderly feels fatigue and bored in assisted walking. Therefore, the aim of our current study is to assist not only physically but also mentally elderly walkers to keep up their motivation. Concretely, we propose a walking promotion method that walkers listen to the beat sound adjusted according to their emotions with the ankle support assistance device. In order to confirm the effectiveness of the method, we conducted two experiments. The first experiment has four conditions where walkers have the device assist or not and the beat sound assist or not. As a result, the gait was improved with the device and sound assist where target cadences were adjusted according to their walking cycle. The second experiment has two conditions where walkers assisted sound adjusted according to their emotions or not. As a result, awakening levels perceived by the walkers remarked higher than ones in usual walking. These results showed the effectiveness of the method and possibility of the control of the emotion while walking.</p>

<p>Many traffic accidents in car driving are associated with sleepiness. It is possible to prevent these traffic accidents by detecting sleepiness of the drivers while driving their car and making them pay attention to driving. Since the possibility that cerebral blood flow is related to sleepiness has been suggested, it is assumed that it would be able to evaluate sleepiness by the changes in cerebral blood flow. In this paper, we measured changes in cerebral blood flow during driving a car using a portable near-infrared spectroscopy device. The results show that oxygenated hemoglobin that is included in the cerebral blood flow tends to decrease with the rise of the sleepiness level evaluated by facial expression and that there is the significant difference in mean oxy-Hb between the two sleepiness level. From this vital reaction, it is suggested that the possibility of sleepiness evaluation of a driver using near-infrared spectroscopy and detection of driver's sleepiness at the level that driver's consciousness become low.</p>

<p>In this research paper, we measured brain activities by near-infrared spectroscopy (NIRS) when a person is feeling the discomfort caused by vibration. In particular, we examined the effect on brain activation by the presence or absence of notice sound at the time of the start of vibration. We used a variance of oxygenated hemoglobin (oxy-Hb) as an evaluation index. We observed a difference in tendency of brain activation between the presence and absence of signal which affect participants in a surprising way. High negative correlation was observed between discomfort to vibration and brain activation around the medial prefrontal cortex, when value of surprise at the start of the vibration is small. High positive correlation is observed between discomfort to vibration and brain activation around the prefrontal cortex outer portion, when value of surprise at the start of the vibration is big. Possibility to evaluate the discomfort is suggested on the basis of brain activation which reflects the changes in the psychological state by surprise.</p>

<p>In recent years, the research in intelligent educational systems has much interest in exploring data from educational settings to understand learners behavior and mental states. Therefore, the means to grasp psychological state of learner has been demanded from a more variety of physiological indicators. In present study, it aims to grasp how the learner approach to the question. By measuring the brain activated by using the multiple-choice questions, this study is to analyze the relationship between confidence and brain activation at the time of answer the question. As a result, in the prefrontal cortex frontal pole vicinity, tended to increase oxyHb when there is high confidence of the answer. In addition, answer time tended to be quick when there is high confidence of the answer. The result shows that it may be able to grasp the state of learner by measuring combination of the change of oxyHb and answer time.</p>

<p>A number of traffic accidents have occurred due to human errors such as inattentive driving and drowsy driving. Sleepiness is considered as one of the causes of these traffic accidents. For this reason, establishment of the sleepiness evaluation method during driving a car has been required. In this paper, we measured changes in cerebral blood flow during driving a car using a portable near-infrared spectroscopy device. The results show that oxygenated hemoglobin tends to decrease with the rise of the sleepiness level evaluated by facial expression and that there are negative correlations between oxygenated hemoglobin and sleepiness level in many trials. This suggests the possibility of sleepiness evaluation of a driver using near-infrared spectroscopy.</p>

<p>高等教育においては，教育サービスの提供者と受給者が目標とする学習成果についての合意形成を行うことで，高い利用価値を共創的に実現することが期待される．本研究では，合意形成の過程を可視化するために，合意された話題の意味内容の概念化を試みる．具体的には，メンターと学習者のカウンセリング事例を取り上げ，サービス提供者，受給者，コンテンツ・チャネルとして発話内容を位置づける．</p>

<p>本研究では，テーマパークにおいて携帯情報端末を用いて混雑情報を取得することができる来場者の満足度を，マルチエージェントシミュレーションを用いてモデル化した．具体的には，期待不一致モデルをもとに，アトラクションに対する効用とアトラクションに並ぶ行列人数とのトレードオフの関係を考慮するようなモデルを構築した．その上で，来場者の満足度向上の観点からテーマパークにおける混雑情報の提供方法を検討する．</p>

Recently, view point of service has been widespread. In learning domains, service providers have been required to adopt it and to provide values co-created by teachers and learners. The value co-creation takes place in joint sphere where providers and customers make interaction, and the value is subjectively perceived by receivers (learners in learning service). That is, a mental state of leaner is one of possible value in the co-creation. The current study aims to develop a conceptual framework which describes mental states of learners for specifying values in the co-creation. As a result, we proposed ontological descriptions of learning event which consists of a teacher, a learner, learning materials, and a learning action by the learner. In concrete terms, we proposed ontological descriptions for receiver states in a learning activity, on the basis of a conceptual model of value co-creation. The ontological descriptions are useful for sharing knowledge on values perceived by learners to design learning services.

We developed a walking assistance apparatus which can be used for neuro-rehabilitation of patients. This apparatus assists only the ankle joints of the equipped person and he/she can raise his/her leg by using the stretch reflex to bi-articular muscle. To confirm the effectiveness of the apparatus, we carried out the experiment for the patients of walking with wearing the apparatus, and the angle variation of the ankle was measured. From the results, by using this apparatus, the dorsiflexion and plantarflexion of the ankle joint of the equipped person can increase while walking. Furthermore, to grasp the lifetime limit of the apparatus, we developed the fatigue test apparatus for the power transmission device of the walking assistance apparatus: from the motor to the worm-helical gearbox through the flexible shaft. This apparatus supplies the power with the motor which is the same as the motor which is used in the walking apparatus. As a result, we could define the limit to replace of the flexible shaft according to the condition of the user.

In recent years, the way to alleviate the disease by using of music that is so-called music therapy, has become popular in Japan. The music therapy is a therapeutic pedagogical technique which is carried out towards the recovery of cognitive and emotional disorders by using of music. Due to awareness of health in the super-aging society of Japan, demands for music therapy is increasing. However, it is difficult to appeal a preference of music for patients who have disorder of speech. Therefore, objective measurement method of emotion, like a measurement method of brain activity, is necessary for provision of music that fit each patients. In this paper, we conducted analysis of brain activity by using of Near-infrared spectoroscopy during listening to music. In our experiment, subjects listened to six kinds of music that combined various tempos and keys. The result shows that each combination of keys and tempos caused different brain activity. It reveals that emotions during listening to music can be identified by using of brain activity.

近年，不気味の谷に代表されるように，人間の外見に精巧に似せて作成された擬人化エージェントに対して，人間は否定的な感情を抱くことが示唆されている．本研究では，擬人化エージェントを観察する時間の長さ，および擬人化エージェントの人間に対する類似度を段階的に変化させ，観察者の情動反応にどのような違いが出るかを，観察中の被験者の生理指標に着目することで実験的に明らかにする．

<p>近年日常的にエージェントが利用されるようになり、今後はエージェントとしてのロボットの普及が加速すると推測される。当該状況下では、人ーエージェント間のインタラクション時の人への負担を軽減することが、普及を促進する上で非常に重要になる。本研究では、人ーエージェント間のインタラクションを負担が軽いと推測される人間のインタラクションに近づける目的で、人が表出する特徴的キューを特定する心理実験を行った。</p>

When a company or an individual disseminate their information on advertising and sales promotion. Between several people who is a client and a designer or a designer and an illustrator and a photographer have a scene to share the visual image. In the transmission communication of visual image, onomatopoeia is able to convey information with a sense of reality. Onomatopoeia because the words that are interpreted by the sense modality. (the sense modality : Visual, Auditory, Olfactory, Gustatory, Tactual) Therefore, in this study, it was intended from association with sensory modality of onomatopoeia in the transfer communication of visual image, and to clarify the relational structure of onomatopoeia. Elucidation of the semantic structure was included ambiguity of onomatopoeia. It will enable the discussion about how to use and understandable meaning of onomatopoeia. The onomatopoeia of 200 words were investigated by the MD method and analyzed using the SS analysis and cluster analysis. And we assume that the semantic interpretation is determined by the strength of the association of the Modal. I discussed the rank structure. As a result, it is considered onomatopoeia of the group that there is a direction of understanding meaning is effective communication to convey the impression with vague. In addition, differences were observed in the technique of description and usage, selection onomatopoeia between expert and novice. It depends on skills and knowledge.

Sharing and managing knowledge on human Kansei obtained in several study fields are of growing significance to providing product's values focused on customers. However, few sufficiently-advanced technologies that realize such knowledge sharing and management have been established. One typical approach is to clarify and systematize concepts as an ontology. In the current study, we developed an ontology to describe correspondence of affective meaning spaces analyzed by semantic differential method. First, we collect literatures reveal the affective meaning evaluated through the five senses and their inter-modality. Second, we described the concepts on the basis of an ontology of color emotions. Then, we demonstrated and evaluated the usefulness of the ontology in a design situation. As a result, we confirmed that the ontology will serve as a useful reference and be of assistance to designers.

This is research on &ldquo;Pause&rdquo; in the speech language producing from a printed book by Transliteration Person: Transliteration Person produces a talking book for visually-impaired people from a printed book. &ldquo;Pause&rdquo; on the speech language is required faithful to original printed book, because Transliteration Person has to reproduce the author's intention. So we experimented in how to use &ldquo;Pause&rdquo; on the speech language with Transliteration Person. As a result, it was shown that &ldquo;Short Pause&rdquo; inserted by Transliteration Person is important to produce faithful talking book from original printed book: &ldquo;Short Pause&rdquo; is 0.5 second or less. In the future work, we would like to show clearly that &ldquo;Short Pause&rdquo; is concerned with comfortable rhythm and tempo for faithful speech language from original printed book.

<p>本研究では、多色配色に対する審美的評価ルールをシミュレートする計算モデルをBPニューラルネットワークで構築した。SD法の心理学実験を通して理論的枠組みであるHelmut Lederの心理学モデルを定量化し、それに基づき計算モデルの構造を決定し、学習率などのシステムパラメータを遺伝的アルゴリズムによって最適化した。性能テストの結果、本計算モデルは多色配色の審美的評価値に対して高い予測能力を有した。</p>

Recently, view point of service has been widespread. In learning domains, service providers have been required to adopt it and to provide values co-created by teachers and learners. The current study aims conceptualization of learners states extracted from utterance contents agreed by a teacher (mentor) and a leaner along with conversation analysis in the previous study. As a result, we proposed ontological descriptions which clarify ASIS/TOBE states perceived by teachers and learners.

<p>学習者の知識・理解状態に加えて心的状態を考慮した学習支援システムの実現のためには，心的状態に関する知識記述・管理が必要である．心理学の分野では，質問紙の評定データと統計モデルに基づいて，学習行為に関わる感情の構造を明らかにする試みがいくつかある．本研究ではそれらの知見を統一的に捉え共有することを目指し，因子分析などの統計モデルを用いて表される感情の次元に関するオントロジー記述を試みる．</p>

Some pauses are found in natural reading a book aloud, which work to support smooth communication. They are supposed to be working in transliterating service for visually impaired people. In this study, we propose types and functions of the pauses in that situation. We conducted an experiment to record speeches of 3 transliterators. As a result, 2 types of pauses were extracted; pause as visual information, and pause as easy-to-understand. In the both type usage, an expert made better use of the short pauses which put 0.5 second or less in the speeches than other transliterators. This suggests that the short pauses have an important function to support the smooth communication. In the future work, we analyze pitches and other parameters to specify the characteristics of expression with the short pauses.

<p>ヒューマンエージェントインタラクションにおいて不気味の谷は重要な課題である．不気味の谷は，人型エージェントに対して人間／非人間の二つの知覚情報処理を行う過程における情報間の齟齬により発生すると考えられるが，この齟齬が不気味さを誘発するメカニズムに関しては統一的な説明がなされていない．本研究では，知覚対象の情動評価に関わる脳部位の機能を定性表現によりモデル化することで，このメカニズムの説明を試みた．</p>

ヒューマンコンピュータインタラクションにおいて不気味の谷は重要な課題である.人間は人間に酷似したエージェントを観察する際に,観察初期の段階では人間を観察する場合と類似した知覚処理を行うが,観察時間が経過するに従って人間を観察する場合とは異なる知覚処理を行うことが実験的に示唆された.このような2つの知覚過程はこれまで感情の二重経路説によってモデル化され,不気味の谷は二つの知覚過程による情報間の齟齬によって引き起こされると解釈されてきた.本研究では,感情の二重経路に加えて海馬の働きに着目することによって,エージェントに対する不気味さ誘発メカニズムのより詳細なモデル化を試みる.

視線計測を用いた先行研究では，四択問題の回答に対して確信がある時の問題読解プロセスは，問題と4つの選択肢を順に走査する規範的なものとなること，確信がない時は走査が乱れることが示唆されている．本研究ではこの知見に基づき，計算機の汎用デバイスからのリソースを用いて多肢選択問題の読解プロセスが規範的か乱れたかを分析することで，回答への確信を推定する手法を考案する．

<p>学習者の知識・理解状態に加えて心的状態を考慮した支援を行う知的メンタリングシステムの実現のためには，心的状態に関する知識記述・管理が必要である．学業に関する文脈で学習者が経験する感情は包括的にAcademic Emotionと呼ばれており，学習者の特性や一時的な心的状態との概念的な関係は明確でない．そこで，知的メンタリングシステムにおける学習者モデルに記述されるべき感情の概念を整理する．</p>

Besides skills in musical expression, exact timing among music players is required in excellent performances. In order to synchronize each respective sound of music among players, they must use a variety of information. In case of four-hand performances on the piano or performances by a band, for example, information such as musical sounds, body movements and gazes is considered to play critical roles. In this study, we empirically investigated roles of the information as communication channels between music players. We conducted an experiment where pairs of players performed music with a constant tempo, one varying its tempo and one including a long rest on the piano. They played the three sets of music both in a face-to-face setting where they could watch each other and a back-to-back setting where they could not. According to the results of the experiment, we then proposed communication models of players in music performances.

自信や興味などの認知，困惑などの情動といった学習者の心的状態に焦点を当てた学習支援を行う場合には，あらかじめそれらの概念が明示されることが望ましい．本研究では，多肢選択問題を回答する学習者の視線と心的状態を例にとり，その知識を記述するためのオントロジーの構築を試みた．

<p>視線は，計算機による学習者の心的状態の理解において有益な指標となることが期待されているが，視線データの分析手法は必ずしも確立していない．本研究では，計算機を用いて学習者が多肢選択問題に回答する場面を想定し，学習者の視線データを取得する実験を実施した．そして，学習者の視線から問題回答プロセスを質的に記述し，問題回答に対する確信度を推定するための分析手法を考案を行う．</p>

<p>近年は，自信や興味などの認知，困惑などの情動といった学習者自身の心理的な側面に焦点を当てた学習者の理解が試みられている．それらに基づいた学習支援を行う場合には，あらかじめ学習者の行動と心理状態に関する概念を明示することが，具体的な支援方略の決定に役立つと考えられる．本研究では，多肢選択問題を回答する学習者の視線と心理状態を例にとり，その知識を記述するためのオントロジーの構築を試みた．</p>

In calligraphy, calligraphers express their emotions in calligraphic characters and that impress people who view the characters. Calligraphers generally acquire tacit techniques for expressing emotions by themselves and, therefore, the knowledge on the technique is difficult to be shared. We experimentally studied features of calligraphic characters expressing five different emotions, using writings created by an expert calligrapher. We classified fundamental strokes comprising the characters based on stroking methods defined in &lsquo;Eijihappo&rsquo; and then extracted knowledge on emotion expression techniques in drawing strokes.

機能指向から感性指向へと消費者にとっての価値が移り変わりつつある現在,膨大なデザイン候補から個人の嗜好に合ったものを抽出することは,情報化社会に置かれた我々の課題と言える.一方で,こうした,情報を高速かつ確実に処理するコンピュータ技術も発展を遂げている.そのため,コンピュータにより人間の感性を反映したシミュレーションを試み,その抽出精度を評価する.コンピュータ上では,人間の脳をモデルとしているニューラルネットワークにより感性評価を行うものとする.結果として,人間の脳活動において存在が示されている感性次元を適用することにより,その学習精度は十分に有効であるが,評価精度については実用レベルに至らないと言える.

色彩感情と色彩の三属性との定量的関係が議論されている．これらの知見は測色量から色彩感情を予測し，コンピュータと人間のインタラクションを円滑にするために有用である．具体的には，システムの知識処理においてコンピュータと人間が概念レベルで知識を共有することが求められる．本研究では，コンピュータと人間が色彩感情に関して，より深い部分で概念を共有することを目的として色彩と感情状態に関するオントロジーの構築を行うと共に，インタラクションに有用であるかを検討する．

色に対する感情は色彩感情と呼ばれ，SD法に代表されるように評定尺度を用いて測定される．SD法および因子分析によって明らかにされる色彩感情の構造は，刺激の評定に用いる評定尺度に依存する．その意味で，評定尺度はそれが表している意味の次元によって評価的(Evaluative)尺度と記述的(Descriptive)尺度に分けられる．本研究ではそれらの尺度同士の関係構造についてオントロジーを用いた表現を試みた．

<p>色刺激に対する印象はSD法および因子分析によってその因子構造が明らかになっている．また，色刺激に対する感情については，感情価，覚醒，支配という三つの次元から解釈されている．印象と感情は互いに深く関連していると考えられるため，それらに関する知見を統一的な視点から理解する必要がある．そこで本研究ではオントロジー工学の方法論を用いて，色刺激に対する印象と感情の関係構造を表現することを試みる．</p>

Human kansei is one of the important factors in recent product development. Thus, management of knowledge about consumers' feelings is an essential task in current industrial design. For the purpose of such the knowledge management, development of a descriptive method to identify feelings is needed. In this paper, we measured implicit attitudes to color combinations using affect misattribution procedure, and we analyzed proportions of positive responses to each stimulus. The results showed that chroma contrast was explained by a concept of perceptual fluency. According to the finding, we discuss significance of the perceptual fluency which links physical features and impressions from the viewpoint of ontological description. The discussion indicated a need for development of a descriptive method for correspondence between physical features and impressions explained by perceptual fluency (e.g. symmetry, contrast, prototypicality).

人間の感性を反映させた製品開発のためには，消費者の印象に関する知識が効率的に管理されることが望ましい．本研究では，対象を知覚する際の行為者の気づきという観点から印象の記述方法を検討する．

本稿では脳という高度な情報処理機構を手本に新しい情報処理機構を構築するという目的の下,逐次分析(SA : Sequential Analysis)を中心として強化学習(RL : Reinforcement Learning),自己組織化マップ(SOM : Self organizing Map)を組み合わせた脳型意思決定モデルを提案する.神経科学や経済学における意思決定に関する先行研究との比較と,本モデルを基にしたコンピューターシミュレーション実験によって,本モデルが人間の意思決定をどの程度再現しているか検証した.その結果から,本モデルの妥当性と,特に大脳新皮質における意思決定処理の再現性が示唆された.

<p>プロダクトやサービスに対して人間が持つ印象は，品質向上に役立てるための重要な知識である．本研究では面接調査によって抽出した印象に関する知識に対して，オントロジー工学の方法論を適用することでその記述を試みる．また，特定の対象世界において印象が言語表現によって観察されることから，印象の記述方法を提案すると共にそれらの区別について考察する．</p>

This study is to constitute a framework for understanding aesthetic sensibility by introducing a mental model called the content space which is defined as a set of all possible contents of consciousness. An aesthetic sensation is a conscious process, thus it belongs to the content space. On the other hand, all mental process must be realized by neural activities which were initially evolved in human ancestors. Because aesthetic sensation is a content of consciousness appeared only in human, it is highly probable that its origin can be explained by a correlating process of biological evolution and cultural development. As the bridge between mental and physiological processes, the emotion and mood relating aesthetic sensation are also considered in the proposed model.

日本の製造業が競争力をつけていくためには，品質，コスト，納期を確保する必要性が指摘されている．製品開発を行う中の上流工程となる開発・設計においては，設計品質の向上および設計力向上を目的として(1) 設計の自動化，(2) 設計ミスの撲滅，(3) 設計資産の再活用といった設計業務の効率化が多くの企業で取り組まれている．(1) 設計の自動化に必要な諸々の数値解析や数値制御のためには，人が設計の中で行っていることを形式知化する必要がある．(2) 設計ミス撲滅に関しては，設計の開始前，途中段階，最終段階のそれぞれにおいて設計者および設計関係者が設計情報のレビューを行いながら，情報の共有を行う必要がある．(3) 設計資産の再活用については，設計資産の蓄積を図りながらデータベースの整備を行う必要がある．これらの取り組みにおいて共通した課題は，人が設計する際に用いる経験やノウハウといった暗黙的な知識を明示的に記述し共有するということであり，それを阻害している最大の要因は，設計に関わる知識が設計者の熟達レベルに強く依存していることである．そこで，本研究では，設計力向上に向けて熟達者が行っている設計方法に関して，設計者が持っている知識とノウハウを表出し，共有，伝承することを目的とした．具体的には，半導体基板の設計を題材として(1) 設計者の観点による設計行為(知識)の表出および設計行為モデルの記述，(2) 設計行為モデルに基づいた設計品質および設計力向上の実務的な評価である．本研究の成果は，設計力向上に向けて熟達者が行っている設計方法とその知識を表出・共有するための設計行為モデルの構築である．このモデルは既往研究で提案されている人間行動モデル CHARM（Convincing Human Action Rationalized Model）に基づいている．既往研究では看護行為を対象にして看護をする動作を扱ったのに対して，本研究の設計行為においては設計者の思考過程を扱った点が異なる．まず，設計者が設計行為を行う際に使用する設計手順書，設計基準書，そして作業マニュアルをもとに設計行為を熟達者の視点から整理することによって，設計の熟達者が何を考えて設計を行っているかということを明示することができた．さらに，構築した設計行為モデルの有用性を検討するために，(1) リファレンスを与えない初心者，(2) 作業マニュアルを与えた初心者，(3) 設計行為モデルを与えた初心者，(4) 熟達者による設計データを比較した．その結果，構築した設計行為モデルは初心者が設計に対する理解を進めるために有用であることが示唆された．ただし，同一の実験参加者の設計経験，参照するリファレンスの情報量の差による影響が明確でないという点が今後の課題として残されている．

画像の特徴量から類似した画像を検索するという，内容に基づいた画像検索（Content-Based Image Retrieval）において，画像の表現と画像の意味内容が直接的に結びつかないことが問題である．画像の表現というlow-levelな特徴量と，画像の意味内容というhigh-levelな概念との間の隔たりはsemantic gapと呼ばれており，その解決のための一つのアプローチとして，画像を構成するオブジェクトに関する概念を記述したオントロジーを利用されている．本研究では，色彩感情に関する概念をオントロジーとしてまとめると共に，色彩感情に基づく画像検索のsemantic gap解決に向けたオントロジーの適用を検討した．オントロジーを利用するアプローチでは，low-levelな特徴量および中間レベルの語彙を構成要素としてhigh-levelな概念が記述された体系を参照モデルとして利用する．例えば，画像における「空」というhigh-levelな概念は，上方領域で青色成分の頻度が高いというlow-levelな特徴量を指していると考えられ，さらに概念と特徴量との中間レベルの粒度では「空間的位置が上かつ色が青」と表すことができる．ここでの「空間的位置」「上」「色」「青」という語彙がオントロジーとして記述されるものであり，概念と特徴量を意味的・内容的に仲介する役割を果たしている．類似した色彩感情を与える画像を検索する際のsemantic gapを解決するためには，画像内のオブジェクトに関するオントロジーではなく，それによって生じる色彩感情に関するオントロジーが必要となる．色彩感情に関する多くの研究成果に共通して，明度はheavy/light，彩度はactive/passive，色相はwarm/coolの反応と対応することが指摘されている．そこで，それらの測定に用いられるSemantic Differential尺度の形容語を取集し，心理的な属性および属性値の概念として定義した．具体的には，明度，彩度，色相という属性概念と明度量，彩度量，色相量という属性値の概念である．これらの概念を定義するにあたり，上位オントロジーであるYAMATO (Yet Another More Advanced Top-level Ontology)を参照した．本研究の成果として，オントロジーで定義した明度量，彩度量，色相量の概念を用いて形容語と色空間座標の関係を明示することができた．画像検索で利用する場合には，ユーザが意図する色彩感情の概念についてオントロジーから形容語と色空間座標の関係を抽出することで，画像検索に用いる特徴量に対して適切な制約を課すことが可能になると考えられる．また，検索結果を出力する際に個々の画像と関連した形容語を提示することも可能であると考えられる．