Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
10326080 | Neural Networks | 2005 | 22 Pages |
Abstract
To study the regularity and complexity of autonomous behavior, the flow of sensory information obtained in autonomous mobile robots under various conditions was analyzed as a complex system. Sensory information time series Xn was collected from a miniature mobile robot during free navigation, and plotted on the return map, the graph of Xn+Ï vs. Xn. The plot exhibited a characteristic trajectory, representing the regularity of the time series. Correlation integral and Lyapunov exponent analysis also showed properties of deterministic chaos; the presence of fractal dimension and positive Lyapunov exponent. Analysis of sensory information obtained in the robot with three different neural controllers revealed that the autonomous robot behaves in such a way that the flow of sensory information is governed by a deterministic rule, and this pattern is unique to each controller. Furthermore, the analysis in various environments exhibited that transitions from one trajectory to another on the return map occur during the course of autonomous behavior. The fractal and Lyapunov dimensions calculated in various conditions indicate that these dimension could be utilized to quantify the complexity of autonomous behavior and the relative difficulty of tasks. Analyses at different evolutionary stage revealed that behavioral performance correlates with fractal dimension. These studies using a miniature mobile robot that allowed to idealize the experimental conditions demonstrated firmly that the complex analysis could be utilized in evaluation and optimization of autonomous systems and the behavior.
Keywords
Related Topics
Physical Sciences and Engineering
Computer Science
Artificial Intelligence
Authors
Md. Monirul Islam, K. Murase,