Robot Fault Detection: At the Analog-Digital Boundary

Abstract

Robots are complex electromechanical systems which often exhibit rapid, unpredictable, and potentially dangerous behavior in the presence of faults. In this paper, we present a novel quantum-inspired approach to fault detection within robot systems. The approach is based on a new fault detection model termed Quantum Analytical Redundancy (QAR) within a Fault Tree framework. The QAR method evolves quantum states with the ability to represent the internal dependencies of data within a provably complete set of fault detection tests. Our research centers on simulation of the approach, with testing and evaluation planned on rigid-link robot manipulator hardware.