Cooperative autonomous mobile robots in a shared workspace

This thesis investigates the problem of control and coordination of multiple autonomous mobile robots in a shared workspace. The overall goal of this study was to develop a method by which two or more autonomous vehicles may work together to complete a given task. In this case the task is to safely and efficiently travel to given goals in an unknown dynamic workspace.;Before the control of multiple autonomous robots can be achieved, the control of a single vehicle must first be completed. Therefore, the initial problem investigated within this thesis is the operation of a single autonomous robot. An autonomous robot must be able to perform map building, localization, path planning, and position control in order to complete a task. In this study, the map building and localization is performed through the use of a remote camera which is interpreted at a PC. Localization is also achieved through the use of motor encoder readings. Path planning is accomplished using a unique hybrid of potential field and fuzzy logic path planning methods. Position control of the robot is achieved using a PI feedback controller.;After the control of a single autonomous mobile robot was achieved, coordination amongst multiple robots was implemented. A centralized control method was used in which the actions of each robot were determined by central PC. The problem of communication between units was solved using a Bluetooth communication network. A unique aspect of this study is that the localization and map building for each unit accomplished by a single remote camera, thus eliminating the need for expensive sensors on each autonomous vehicle.;The results of this study show that the unique path planning method offers a marked improvement over the traditional potential field path planning method. Using the hybrid potential field and fuzzy logic path planning method, the vehicle is able to reach its goal safely and more efficiently. The findings of this study also demonstrate that it is possible to control multiple autonomous vehicles using one shared remote vision system for localization and map building. The goal of controlling multiple autonomous mobile robots was accomplished without the use of expensive sensors at each vehicle.