A control architecture for intelligent robots based on graph manipulation for planning, monitoring, execution, intention switching, and fault recovery: Dissertation

A new robot control architecture is proposed which combines the functions of planning, intention switching, execution, plan validity and plan execution monitoring, and fault recovery. Emphasis is placed on providing a technological basis for building control systems for intelligent robots by modeling these functionalities using simple, though powerful, graph manipulation techniques and introducing different types of plan components for monitoring, planning, and execution using an object-oriented design approach. The proposed architecture is related to existing approaches and results obtained from the research areas of monitoring and planning. An implementation of the architecture is described with results from tests in a simulated environment being reported. A new method for grasp and motion planning based on hierarchical fuzzy logic rule sets is presented and its feasibility is verified by applying it to an industrial paper roll manipulator.