IN RECENT years, the need for a solution to the failure detection, identification, and evaluation (FDIE) problem for aerospace vehicles that includes all subsystems over the entire flight envelope has been widely acknowledged [1–4] and has become a major objective of NASA’s Aviation Safety Program [5]. Previous methods, including state estimation or observer-based schemes [6–8] relying on Kalman or other filters and artificial neural-network-based schemes [9, 10], have focused on specific failures and limited areas of the flight envelope. The complexity and extremely high dimensionality of the problem of detecting various aircraft subsystem failures over the entire flight envelope require adequate tools. Recently, a new concept inspired from the biological immune system was proposed for aerospace systems failure detection [11, 12]. The artificial immune system (AIS)-based fault detection operates in a similar manner, as does its biological counterpart (according to the principle of self/nonself discrimination) when it distinguishes between entities that belong to the organism and entities that do not. This paradigm can potentially directly address the complexity and multidimensionality of aircraft dynamic response in the context of abnormal conditions and can provide the tools necessary for a comprehensive/integrated solution to the FDIE problem. A critical issue for the AIS-based detection is the generation of adequate detectors [13]: that is, the definition of regions in the hyperspace of relevant parameters (identifiers) that are only reached when abnormal conditions are present. To date, there is no deterministic method to perform this task and available algorithms rely on random location of detectors and search for uncovered regions. In addition, the need to computationally optimize the detector set for online detection and to ensure maximum coverage of the self/nonself without overlapping for good detection performance makes

revision received 22 September 2009; accepted for publication 16 October 2009. Copyright© 2009 by the authors. Published by the American Institute of Aeronautics and Astronautics, Inc., with permission. Copies of this paper may be made for personal or internal use, on condition that the copier pay the $10.00 per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923; include the code 0731-5090/10 and $10.00 in correspondence with the CCC.∗ Graduate Student, Department of Mechanical and Aerospace