Statement of the Problem:

Design, implement, and test a real time system using computer vision and machine learning to demonstrate learning behavior in a miniature mobile robot.

Goal of the Project:

The goal of our project is to employ a miniature robot which, through environmental sensing or computer vision or both, will learn to navigate a maze by the optimum route. Through design, fabrication, implementation and testing, we will be able to investigate, evaluate, and implement various theoretical concepts including robot design, maze design, interface design, learning algorithms, and real-time system engineering.

Overview of the Design Solution:

Design and create a test maze with moveable obstacles, which will allow for rapid changes in the maze design. Employ a miniature robot equipped with proximity sensors and a radio frequency robot/computer interface. In one scenario, we will employ an "eye in the sky" video camera to determine the position of the robot in the maze and then employ this information within the learning algorithm for robotic navigation in real time. In a second scenario, we will employ the robot's onboard sensors solely along with the learning algorithm to discover the optimum route, which the robot must take to travel from its initial position to the goal position in the maze.

Mark Patterson 6/25/99

Introduction | Overview | Maze | Interface | Learning | Vision | Robot | Integration | Bibliography