Special Article of Chaos, Self-Organized System
Professor, Faculty of Engineering, Hokkaido University, Kita-13, Nishi-8 Sapporo, Hokkaido 060 Japan
The actual environments under which robots are going to operate from now on are complex and sometimes unstable unlike the arranged environments in factories. It is becoming increasingly necessary for the robots to be able to cope with complexities by maintaining a symbiotic relationship with man who are behaving in mental world in various manners and life styles. This is not a task for the distant future but has already been posing daily problems in the fields of computer information communications that reach an international networked society. In this world, there is already a limitation to top-down controls based on hierarchical knowledge and guidance instructions. Rather, the situations indicate that this world should be taken as a complex adaptive system that requires recognition of the environments based on an autonomy of the system, adaptation and learning by behavior, and formation of orders according to self-organization, through a bottom-up approach. This is a considerable task to be tackled from now on if it is desired that robots are to play an active role in various parts of society in the near future. In considering this problem, it is important to learn from nature, ecological systems, and life systems. In recent years, new academic research fields have been generated that imitate the information processing functions of creatures such as recognition, evolution and adaptation, then reproduce these information processing functions in CG’s and robots, and apply them to science and technology. This has been posing large topics relating to the autonomy, adaptation, learning and evolution of complex systems and finally to the creativity of a living body system. The key concept in this case is a chaos edge system called by Langton. In a complex system like a life, there is a self-organization marginal level of a certain scale, above which information is scattered, and below which information is fixed like a crystal. An integration is progressed and emerged in these systems. This characteristic is called the chaos edge system. Along with the outbreak of research in this new field, this has given substantial influence to the research into conventional fuzzy theories, neural networks, and intelligence systems such as artificial brains. Furthermore, along with the development of calculation methods called genetic algorithms which have been learned from the adaptive evolution of creatures, chaos, fuzzy, neuro, AI and GA are all directed towards the structuring of a new intelligent system. We cannot take our attention off the research in this new field. In order for robots to maintain a symbiotic relationship with man in daily life, the new development relating to the complexities of the creatures as described above is necessary. We already open every year an intelligent system symposium which aims at the merging and new developments of fuzzy, neuro and AI. At last year’s FAN’95, we opened an organized session whose theme was chaos and self-organization systems. This special article has been organized now to high-light the papers announced at this OS.It has been possible to take up a wide range of issues including ecological systems, living systems, chaos, robots and the mind. I would like to express my thanks to those who have found time in their busy lives to contributed their papers to this special article . There are many more papers that were presented at the symposium and only a function of them were included in this article because of the limitation of space. I look forward to organizing a similar article again for this book, and it will be my great pleasure if readers can understand the latest situation of the research in this field from this issue. Finally, my thanks are also due to Dr. Tetsuro Yabuta (NTT) and Dr. Tadashi Iokibe (Meidensha) who helped in editing this article.