Designing a Robotic System Implementing Autonomous Path-finding, Image Processing, and Mechanical Actuation to Clean and Organize a Space

Abstract

Robotics and autonomous systems occupy an ever increasing presence in our daily life. This thesis attempts to explore and develop a robotic system that combines material from a wide variety of topics including Autonomous Path-finding, Image Processing, and Mechanical Actuation. The idea is to enable a robot to clean a room autonomously by retrieving objects that lie on the ground around it. The design utilizes TurtleBot, ROS (Robotic Operating System), Xbox Kinect, Python, and Arduino in order to accomplish this goal. The project was split into three tasks: creating the vision algorithm used to detect small objects in front of the TurtleBot, designing and constructing the linear actuator used to lift the detected objects, and utilizing python code to enable autonomous driving. While a fascinating project in itself, this thesis aims to identify successful aspects of the robotic system design in order to allow the reader to use the knowledge to create far more sophisticated and accurate robotics systems in the future.