'Point-and-click' camera-space manipulation, mobile camera-space manipulation, and some fundamental issues regarding the control of robots using vision

This dissertation presents three distinct, but related projects tied together by the vision-guided method of controlling robots named camera-space manipulation (CSM). The first project consists of developing a "point-and-click," semi-autonomous version of CSM. This version of CSM overcomes many of the difficulties associated with the visual recognition problem inherent in any fully autonomous vision-guided method of control. Equipped with a graphical user interface, "point-and-click" CSM is easy for an operator to use and facilitates the changing of different types of robotic tasks. In the second project, a detailed comparison of CSM to a competing method of control, visual serving, is made in order to ascertain the relative strengths and weaknesses of the methods. In the context of a two-degree-of-freedom example problem, a point of convergence is reached between the two methods providing a good basis for comparison. Finally, in the third project CSM is adapted to control both the nonholonomic and holonomic degrees-of-freedom of a rover-type vehicle equipped with a robotic arm. This new method of control, mobile camera-space manipulation (MCSM), enables the precise positioning of a tool mounted in the robotic arm at some target juncture located in the environment of the rover using the minimal number of degrees-of-freedom.