Design of Obstacle-Avoidance and Object-Tracking Controllers for Soccer Robots Based on Artificial Immune Network
| Autor: | Chin-Chun Yu, 游智鈞 |
|---|---|
| Rok vydání: | 2003 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 91 In this research, a behavior-based controller is designed for a soccer robot by using the concept of artificial immune network (AIN). This design involves four procedures, namely, planning a way to detect the invading antigen, defining the antigenic determinant and antibody combining-site, constructing a mathematical model for adaptive mechanisms which active or suppress lymphocytes, and devising robot actions to the invasion of antigens. These design procedures are confirmed by computer simulation before applied to robot behavior control. The AIN behavior-based controller is equipped with servo functions including sensing, learning and adaptation, and action, so it will ensure the robot to attain the behavior of obstacle avoidance and object-tracking. A soccer robot of RoboCup small-size league is carried out in this research. The image of the surrounding environment is acquired by a panoramic vision system on the robot, and is fed back to the controller via a wireless communication system. The panoramic vision system is composed of a convex mirror and a CCD camera. The mirror is placed on the top of the robot, and is inclined to the moving direction with a small angle. Therefore, the mirror can reflect more frontal image to the CDD camera. The acquired image is distorted and recovered by a method based on TSK fuzzy system. The developed AIN behavior-based controller was applied to the small-size robot to verify the function of obstacle avoidance and object tracking. The results show that the robot conforms the behavior-based controller, and has accomplished the planned tasks. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|