リサーチユニット総覧(Research Unit Magazine)

ソフトコンピューティングリサーチユニットProblem-Solving by Approaches with Tolerance of Ambiguity/Fuzziness

代表者:鬼沢 武久    中核研究者:安信 誠二  宮本 定明  
他のメンバー:佐藤 忍  川村 洋平  庄司 学  延原 肇  星野 准一  澁谷 長史  白川 友紀  遠藤 靖典  

野田五十樹 Vicenc Torra Rahmat Widyanto Bede Barnabas 










 日本とインドネシアは世界有数の地震発生国ですので、2国が協力して地震発生後の人的被害状況の時空間的な把握に力を入れています。特に、地震が起こって通信網が遮断された際に省バッテリーの通信装置ZigBee*1 (ジグビー:図2)をいくつかつなげ、ソフトコンピューティングを用いて被害を最小限にくい止めるといった救命戦略立案、例えば非常用食料、水などをどれだけどこに送れば良いのかの立案、を行っています。また、動的な人的被害の推定問題として定式化を行い、被害状況の把握をするためのアプローチ方法を開発しています。





● 共同研究:知的設計器による回路設計支援システムに関する研究
● 共同研究:オンライン辞書を利用した学習支援に関する研究
● the 5th Int’l Symposium on Computational Intelligence and Industrial Applicationsの企画
● 第8回筑波大学ソフトコンピューティングセミナー開催




Problem-Solving by Approaches with Tolerance of Ambiguity/Fuzziness


Unit representative: Professor Takehisa Onisawa, Faculty of Engineering, Information and Systems

Unit members: 15 (11 faculty members, 0 postdoctoral fellows, 4 from other organizations)

Key words: Soft computing, intelligent information processing, communication engineering, disaster control, geo-sensing



  The research unit “Soft Computing” aims to serve as a research basis for intelligent information processing systems based on soft computing, and consists of researcher members from a variety of fields including clustering, machine/reinforcement learning, data analysis, large-scale multi-media information processing, analyses of entrance examinations, geo-sensing, intelligent control, and Kansei information processing. As an organized research institution, our research unit tries to develop new and more essential frameworks for research problems. “Soft computing” is a new research field designed to integrate findings from a variety of research fields to address problems that cannot be easily solved by existing conventional methods.

Although soft computing has soft in its name this is quite different from the field of software research


Soft computing: 1) is tolerant of ambiguity/fuzziness in solving research problems, and 2) is based on the idea that different approaches are used while facilitating their collaboration by compensating for the disadvantages of one approach with the advantages of another. Since it is very difficult to accurately process every piece of information in real life, the soft computing approach sacrifices the accuracy of information to some extent, and tolerates of ambiguity/fuzziness to address problems that cannot be solved by a crisp approach using only 0 and 1. The framework of soft computing combines the advantages of multiple methods to solve problems that cannot be solved by a single method (Figure 1). Soft computing is designed to obtain results in the almost same manner as humans do using their senses.

Figure 1: Soft computing

Figure 1: Soft computing




Development of emergency strategies using ZigBee and soft computing

As both Japan and Indonesia are known as earthquake-prone countries in the world, they are cooperating with each other to estimate the status of human damage following an earthquake in terms of time and space. In particular, the research unit tries to develop emergency strategies to minimize damage in the case of the failure of communication systems due to earthquakes, whereby multiple ZigBees*1 (Figure 2), low-battery-consumption communication equipment, are connected with each other based on soft computing, which provides information on the amounts of emergency provisions and water to be distributed and their destinations. The research unit is also involved in the development of approaches for estimation problems such as the recognition of damage, and the formulation of dynamic human damage.

Figure 2: Zigbee

Figure 2: Zigbee

*1: A short-range wireless communication standard. The communication distance is 100 meters.

Social contributions and achievements
● Collaborative study: Study on circuit design support systems using intelligent design equipment
● Collaborative study: Study on learning support using online dictionaries
● The 5th International Symposium on Computational Intelligence and Industrial Applications
● The 8th Soft Computing Seminar in the University of Tsukuba

 (Interviewed on July 31 , 2013)