single-rb.php

JRM Vol.10 No.4 p. 283
doi: 10.20965/jrm.1998.p0283
(1998)

Editorial:

Special Issue on Complex Systems in Robotics

Sadayoshi Mikami, and Mitsuo Wada

Chaos, Complex Systems Engineering, Faculty of Eng.Hokkaido University, N-13, W-8, Sapporo 060-8628, Japan

Published:
August 20, 1998

The Really “intelligent” robots predicted by science fiction have yet to appear, and robotics research seems to have reached a wall in dealing with the real-world environment. The robot is a unique device that it interfaces directly with the environments, including humans, machines, and nature. The world is very complex and changes dynamically. Robotic research must thus consider how to deal with such dynamcal complex world by means of machines. Our special issues on the complex systems in robotics introduce current representative approaches and attempts to answer these questions. The approach from a complex system point of view deals with new directions in robotics, for the above reasons and provides ways to view things dynamically, in a way that goes beyond traditional static control laws and rules. As these issues show approaches are divergent and ongoing. Modeling and forecasting the world is not haphazard. If requires direction. Even robots that navigate traffic, for example, must have a model to forecast unknown dynamics. Human interfacing requires far more difficult approaches than we take now. Recent developments in theory of chaos and non-linear predictions are expected to provide ways to enable these approaches. Robot interaction with the environment is one of the fundamental characteristics robots, and any interaction incorporates underlying dynamics; even robot-to-robot interaction exhibits deterministic dynamics. We will see how to deal with such complex phenomena through the articles predicting chaotic time series in these issues. Very rapid adaptation to the world is another way of coping using a brute-force approach. Reinforcement learning is a promising tool for working in a complex unknown environment. Learning robots affect both their environment and other robots. This is the situation in which we must think of the emergence of complexity. This may provide a rich source of possible tasks, and we must consider its dynamic nature of it. Many interesting phenomena are shown in the papers we present, applying reinforcement learning in multi-robots, for example. Finding good solutions wherever possible is a rather static solution but must incorporate the mechanism of how nature generates complexities and rich variations. Evolutionary methods, which many papers deal with in this issue, involves trends in complex systems sciences. Robotics applications must consider practical achievements such as rapidity, robustness, and appropriateness for specific applications. These issues provide a variety of robots and automation problems. Of course there are lots of other ways for this quite new approach and it should be worth cultivating because it is just the way we expect that robots should go. These special issues are organized from many papers submitted by researchers, all of whom we thank for their contributions. We hope these issues will help readers to familiarize themselves with the many trends in researches beyond engineering approaches and treat their practical implementation. This area is now very active, and we hope to see many papers related to this theme submitted to this journal in future.

Cite this article as:
Sadayoshi Mikami and Mitsuo Wada, “Special Issue on Complex Systems in Robotics,” J. Robot. Mechatron., Vol.10, No.4, p. 283, 1998.
Data files:

*This site is desgined based on HTML5 and CSS3 for modern browsers, e.g. Chrome, Firefox, Safari, Edge, Opera.

Last updated on Jun. 19, 2021