"Reconfiguration Algorithms for Failure Mitigation Electronics"

Sharon S. Graves1 and Donald E. Harper2
1NASA Langley Research Center
2SWALES Aerospace

Abstract

Aerospace prognostics and failure mitigation electronics use embedded self-recovery mechanisms encompassing: built-in test, diagnostics, prognostics (or predictive diagnostics), health monitoring and failure mitigation. Embedded self-recovery mechanisms implement health management at the device and circuit level rather than at the system level.  The main advantage of embedded self-recovery is that small deviations can be compensated for locally, providing a first, and rapid mechanism of self-defense, before the deviations/degradations create an actual fault which is propagated to higher levels.  

This paper comprises the reconfiguration algorithms and modeling for electronics that allow for quick and efficient development of smaller more robust systems by providing a means of adapting to changing environments and capabilities.  The diagnostic electronics has general applications to areas such as robotics, flight control, data mining, high-end computing, instrument/industrial control, and many others.  Diagnostic reasoning algorithms operating in a RISC processor core embedded in a Field Programmable Gate Array perform rapid real-time classification and identification of faults, failures, and damage when they occur.  The diagnostic and failure mitigation software collects data from sensors, processes the data, determines the current and evolving health/damage state and mitigates the failure and damage. A hierarchical approach is used for health/damage state determination that includes signal processing methods for feature extraction, anomaly detection, analytical tools, and artificial intelligence methods including evolutionary reconfiguration algorithms.  A model based design approach provides quick-turnaround for development and includes hardware-in-the loop testing (see Figure 1) of the electronics.

Figure 1 Elements of embedded diagnostics and failure mitigation

 

2006 MAPLD International Conference Home Page