Research Title Basic research about intracranial higher cognitive functions Research Staff Research representative：Yinlai JIANG Colleague：Tatsuhisa TAKAHASI( Asahikawa Medical College), Masanaga IKENOE(Asahikawa Medical College),Hirotaka YANAKIDA( Yamagata University), Qizhang PIAO(Kochi examination clinic) Research Background Human have the ability to reason and recognize the figure and word which have missed a part. This kind of recognition of scrappy visual information is called as “visual interpolation”, it is the important feature of intracranial visual information processing. This research, using the deficient word as the signal, investigate the mechanism of visual interpolation. By developing the early detection method for dementia based on deficient words, we do the contribution to nursing prevention of dementia elderly. The population of dementia elderly is estimated to be 5.5 million in 2025, their caring and therapy is the country problem. To address this problem, this kind of early detection is most effective. Although, the current diagnostic method (Revision: Hasegawa-style Mini-Mental State Examination HDS-R), is also useful to diagnosis the dementia, however, it is not effective to detection the early patient and pre-symptomatic patient. From now on, we do the research about the “quantification method of perception about the word missed a part”, discover the quantitative relationship between the declines in the perception ability on the word missed a part and aging. This research could apply this method to a lot of subjects, could statistically and quantitatively clarify the relationship between the perception on the word missed a part and asymptomatically lesion of brain being to dementia symptoms. Based on this, we develop the method of early detecting the perception which could be easily conducted on the group medical examination or in-home.

Research Title Systematization of distance type fuzzy reasoning method Research Staff Research representative：Shuoyu WANG Colleague：MIZUMOTO Masaharu(Osaka Electro-Communication University), TSUCHIYA Takeshi (Hokkaido University of Science) Research Background The phenome of misunderstand that “Because of being fuzzy, properly doing is ok”, was generally find in the researchers who doing the research about fuzzy. This is not only the problem of fuzzy set theory, the previous [adaptation inference method] is also pointed out have this problem. On the other side, in the binary logic world, we have the philosophy that “do our best to be strictness”. However, strictness and minuteness have the completely different concepts, that is, strictness isn't necessarily with a minuteness results. For example, the binary logic, the elderly is strictly defined as the people over 65 years old, the people whose 65th birth day is later one day than now is not the elderly, and the next day could be seen as the elderly, this is quite unnatural, which cannot minutely represent the objectivity facts. Our research standpoint is to develop the more precisely expression than the fuzzy concept (ambiguous concepts) and the reasoning method to get the more precise reasoning results. Therefore, the formula to precisely calculate the distance between the fuzzy sets was designed, the base of reasoning to reason the distance between the fuzzy set, basic algorithm of distance-type fuzzy reasoning method as the beginning, from many aspects such as: the numerical truth-value and language truth-value, one rule and many rules, learning and association of ideas, continuity and discrete, direct space and special space and so on, new reasoning algorithm was serially proposed, many kinds of special features was mathematically proved. Human reasoning engineering was developed by compiled by these algorithms. This method have already been used in the nonlinear control, modeling, web searching system, automatic music selection, guarding robot and so on, and its effectiveness has already been conducted.

Research Title Design method of biological information feedback control system Research Staff Research representative：Shuoyu WANG Colleague： Research Background The current control engineering discuss about the transfer of signal. For example, we use the control variable: position, speed, voltage, current, and temperature, pressure etc. which could be represented by physical quantity of measuring as the control subject. In mathematical, it could be described by one of the element in the set of representing concept. That is, current control engineering should be understood as the education that influence the element in set. But, as the example shown in the following figure, in the practice society, the concept expressed by set, and there are a lot of construction of influencing these thing, it is impotent. We called this system as the bio-feedback system. And, not only the signal, the object also movement. To make the object movement, the energy is necessary. Therefore, this research is to construct the theory system of new control engineering based on the energy and concept set.

Research Title Rubber string type associative memory algorithm Research Staff Research representative：Shuoyu WANG Colleague：MIZUMOTO Masaharu(Osaka Electro-Communication University), TSUCHIYA Takeshi (Hokkaido University of Science) Research Background Human have the case to associate friend’s character by the past thing such as their past photos, sometimes, we associate the character of our friends by the present received from them. And we may associate the scenes that we playing together with our friends by the same photos. That is, the associative memory of human have the robust. This research develop the associative memory algorithm with human robust to the proposed rubber string type associative memory model. Basic consideration method is shown as following figure, the output prototype pattern space, mass of ball is thought to take from the mass point to hypothetic rubber by pulling. Tractive force of rubber, in the input prototype pattern space, the distance between the practice input and each input pattern was set as the counter example. The associative memory utilizing the designed dynamic modeling have the following characters. （１）Theoretically graduate the stability of associative memory. （２）Since with the dynamics, it is possible to imitate the associate memory of brain in true meaning.

Research Title High-speed learning algorithm to freely define the learning error Research Staff Research representative：Shuoyu WANG Colleague：MIZUMOTO Masaharu(Osaka Electro-Communication University), TSUCHIYA Takeshi (Hokkaido University of Science) Research Background 我々は素晴らしい学習能力を持っていればこそ、物質財富だけでなく精神財富も含めて継承でき、人類社会が少ない無駄で今日まで進歩してきたのである。学習のキーは学習時間と誤差である。今まで提案されている学習アルゴリズムの殆どは学習誤差と時間のうちどれかを見失ってしまっている。本研究では、人間の学習行為から考察することで、独自な学習アルゴリズムを開発している。 一言で学習といってもきわめて広い概念であり、各分野では自分なりの解釈があり、また個人によっても学習に対しての認識も様々である。ここでは、学習とは少なくとも次の三つの機能を備えることが必要であると考えている。すなわち １）環境からのデータより新知識の発見、理解そして正確に吸収する機能　 ２）もし既在の知識集合中にある部分集合が変化している環境に適合できなければ、古い知識として見なして速やかに更新する機能。　３）何らかの原因で間違った知識を認識し、直ちに修正する機能。 例題として、非線形レーザー工学モデルを教師データとして、学習した結果は下図に示す。

Research Title Creativity cultivation method by combining the robot and comic Research Staff Research representative：王 碩玉 Colleague：牧野 圭一（京都精華大学 マンガ学科）・ 河田 耕一（高知工科大学 工学部） 　　　　　　　　井上 喜雄（高知工科大学 工学部）・ 土谷 武士　（北海道工業大学） 　　　　　　　　 Research Background 創造力は、豊かな感性から生れ、空想できる空間で高められ、そして現実世界で確められるものである。感性は現実の世界からしか得られない。よって、創造性育成は図に示すように空想と現実の間で行うべきであると考えている、図に示す流れを「創造性育成の螺旋状サイクル」と呼ぶ。本研究では、創造性育成の螺旋状サイクルにしたがってロボットの実体性とマンガの空想性を利用して、小学生を対象とする創造性教育法を開発することを目的としている。

Research Title Path generation and trajectory planning method of the intelligent model robot Research Staff Research representative：王 碩玉 Colleague：尚涛（中国、瀋陽工業大学） Research Background ロボットの経路生成と軌道計画は、研究問題として提起されてから数多くの優れた研究成果があげられた。しかし、 人間の車運転のように自由自在に実環境を走ることができる自律型ロボットはまだ実現できていない。本研究では、運転手の知能を模倣することにより、知能ロボットの自律的経路生成と軌道計画を実現することを目指している。運転手の行動知能を獲得するために、シミュレーションシステムを開発し、運転手の行動知能の抽出法を開発している。

Research Title Development of family doctor system Research Staff Research representative：Shuoyu WANG Colleague：Xia YU（Shenyang University of Technology）・邸紅香（瀋陽医学院）・片岡万里（Nursing Course of Kochi Medical School） 　　　　　　　　石田健司（Nursing Course of Kochi Medical School） Research Background もし気軽に相談できる保健関係の先生が傍に居れば、何時でも何処でも健康チェックや健康助言を受けられ、日頃の健康づくりや疾病の早期発見・早期対応ができ、自覚症状を無視したため手遅れになることに後悔する人々が無くなり、何方も健康かつ長寿で社会に貢献できる。しかし、現状では専門知識を持っている保健関係の先生が不足しており、全ての人に一人の先生が付くことは到底不可能である。したがって、我々が、専門知識に基づいてエキスパートシステムを構築することを考え、具体的に健康状態のチェックや健康増進など機能を備えるソフトウェア?「ホームドクターシステム」を開発している。ホームドクターシステムを完成すれば、何方もパソコンで何時でも何処でも健康チェックと健康助言を受けられ、ユビキタス健康増進を実現できる。

距離型ファジィ推論法・連想記憶・学習アルゴリズム、距離型ファジィ推論法の理論構築等を行っています。