Dynamic Fault Reconfigurable Intelligent Control Architectures for Robotics

Abstract

In this paper we describe new progress in our development of an Intelligent Control Framework for robots which dynamically reconfigures itself to cope with faults in either sensors or joint hardware. The Framework is configured to allow the incorporation of new approaches for on-line critiquing of user plans and commands within the framework. We discuss integration of the two components above to produce an Intelligent Robot Operating System which can tolerate failures or unexpected actions from both the logical (user) world and the physical (manipulator) world and continue operation where possible.