| In this work, various parts of an autonomous ground vehicle (AGV) for agricultural purposes are designed. This AGV is part of a larger project that, in a cooperation with an unmanned aerial vehicle (UAV), can detect and remove infected potatoes plants by virus Y (PVY). In the first part of this work, the design of a prototype chassis for the AGV was analyzed. This prototype is a four-wheel powered vehicle. An optimization routine was used for finding the ideal size and material for the chassis. In the second and third parts, the design, implementation, and performance of autonomous navigation and obstacle avoidance systems for this AGV are discussed. To reach the maximum accuracy for navigation, an RTK GPS has been used. Also, for obstacle avoidance in this AGV, a simple 2D LIDAR sensor was installed and tested. The fourth part is related to detect the exact location of the infected plant by an advanced robot vision system and proper image processing algorithm. This system includes two RGB cameras and utilizes the HSL format of images for image processing. The fifth part of this work is related to design of an efficient end of arm tooling (EOAT). The most appropriate mechanism for complete removing infected plant was designed and a prototype model of it was built and tested after passing the motion analysis in a CAD model. In the last part of this work, a novel singularity analysis for robotic mechanisms has been presented, and the idea is developed by an example of a spatial mechanism (RRSS). The outcomeof this work includes design and implementation of various parts of an agricultural ground vehicle. The motion study and stress analysis for the mechanical sections of this vehicle are analyzed in SolidWorks, and a prototype model of each part is tested in the field. All electronic sections (including various GPS, sensors, cameras, motors, etc.) have been checked
xvi| P a g eTAHER DEEMYADPh.D. Dissertation Advisors: Prof. Alba Perez Gracia & Prof. Anish Sebastianone by one in different situations based on function. Also, codes and algorithm for image processing and navigation were tested in various conditions. KeyWords: Optimization, Stress Analysis, Chassis Design, Autonomous Vehicle, Four Wheel Drive Vehicle, CAD Stress Analysis, Mobile Manipulator, Roguing Mechanism, Agricultural Robot, End of Arm Tooling, Autonomous Virus Removal, Navigation, RTK GPS, Obstacle Avoidance, Agricultural Robot, Autonomous Ground Vehicle, Pixhawk, Robot Vision System, HSL format, Image Processing Algorithm, 2D object detection, Image Segmentation. |