An Embedded Intelligent Mobile Robot Based on SOPC

• Main Authors: Wen-zong Tung, 童文宗
• Other Authors: Kau-shing Hwang
• Format: Others
• Language: zh-TW
• Published: 2007
• Online Access: http://ndltd.ncl.edu.tw/handle/65277660013756884280

碩士 === 國立中正大學 === 電機工程所 === 96 === This thesis develops methodologies and techniques for designing and implementing an embedded intelligent mobile robot based on SOPC. Such a system can be operated to any location according to users’ commands. The user are able to view real-time images and steer the vehicle at the same time using a standard web browser on any local or remote computer on network. Another function is to let the robot achieve seeking goal automatically, so we have add the capability of performing of robot learn in the system. To design appropriate actions of mobile robots, designers usually observe the sensory signals on the robots and decide actions from the viewpoint of some desired purposes. This approach needs deliberative consideration and abundant knowledge on robotics for a variety of situations. We proposed the Reinforcement Learning-based Decision Tree Algorithm to solve these problems. In the method, every robot is a learner, and the designers only need to demonstrate behavior to robot, which then can record these information between actions and environmental states and induce the relationship between the sensory signals and the actions. The system is designed in the bottom-up approach. Starting from the low-level hardware system design, the work completes the customized SOPC design on image acquisition system built with CMOS image sensor and integration of the wheeled mobile platform. The top level software application designs of the system include the construction of the Operating System environment, the image fetching implemented by HW/SW Co-design, and an interactive web-based user interface implemented by CGI programs. All the proposed functions have been verified and have been highly integrated into a complete embedded system.