A n-dimensional convex hull approach for fault detection and mitigation for high degree of freedom robots humanoid robots

摘要

This work shows a plan for error state determination, diagnosis, and mitigation using autonomic computing tools and tecniques on the Hubo robot. An n-dimensional geometric enclosure is constructed from periodic measurements of the robot's normal operating state. Similar hulls will be constructed for unique faults encountered during runtime. A mapping of these faults to applicable mitigations will be dynamically constructed and will aid in mitigation selection. The successful application of the mitigation will bring the robot back to a safe operating state.