The development of an interactive synthesis tool for intelligent controllers of modular, reconfigurable robots

Intelligent controller design for modular, reconfigurable robot architectures was investigated. In robotic systems available today, the controller is the main limiting factor for enhanced performance. Numerous researchers are studying new controller architectures to allow for the implementation of advanced software, but few have targeted their research at reconfigurable designs. This work has focused on the development of a modular controller architecture that can be reconfigured to meet the changing needs of new manipulators and new tasks. The dissertation work was divided into two parallel lines of research. First, a roadmap assessing the required technology for an intelligent controller was created. A complete model of robotic manipulators was developed which included the arm dynamics and the dynamics of the joint actuator systems. This model was used to determine the desired update rate for command signals from the robot controller. Second, a synthesis package was developed on a Silicon Graphics workstation to simulate controller architectures. This package allows the user to assemble computer components and select controller algorithms, and then evaluates the effectiveness of the design. Bottlenecks are easily identified and improved architectures can be rapidly created. This synthesis package was used to underscore the issues discussed in the roadmap.