Task Planning for Robotic Manipulation in Space Applications

Space-based robotic systems will require new technologies of planning and manipulation to accomplish complex tasks such as diagnosis, repair, and assembly. A review is presented of results of work in assembly task representation, discrete task planning, and control synthesis which provide a design environment for assembly systems, and which extend to planning of manipulation operations in unstructured environments. In this approach, assembly planning is carried out using the AND/OR graph representation which encompasses all possible partial orders of operations and may be used to plan assembly sequences. A new algorithm for planning disassembly and repair using the AND/OR graph is introduced, and examples of repair sequences generated for a satellite electrical module are described. For discrete task planning, the configuration map facilitates search over discrete parameters in the space of bounded configuration sets.