Development and evaluation of a fault detection and identification scheme for the WVU YF-22 UAV using the artificial immune system approach

Abstract

A failure detection and identification (FDI) scheme is developed for a small remotely controlled jet aircraft based on the Artificial Immune System (AIS) paradigm. Pilot-in-the-loop flight data are used to develop and test a scheme capable of identifying known and unknown aircraft actuator and sensor failures. Negative selection is used as the main mechanism for self/non-self definition; however, an alternative approach using positive selection to enhance performance is also presented. Tested failures include aileron and stabilator locked at trim and angular rate sensor bias. Hyper-spheres are chosen to represent detectors. Different definitions of distance for the matching rules are applied and their effect on the behavior of hyper-bodies is discussed. All the steps involved in the creation of the scheme are presented including design selections embedded in the different algorithms applied to generate the detectors set. The evaluation of the scheme is performed in terms of detection rate, false alarms, and detection time for normal conditions and upset conditions. The proposed detection scheme achieves good detection performance for all flight conditions considered. This approach proves promising potential to cope with the multidimensional characteristics of integrated/comprehensive detection for aircraft subsystem failures. A preliminary performance comparison between an AIS based FDI scheme and a Neural Network and Floating Threshold based one is presented including groundwork on assessing possible improvements on pilot situational awareness aided by FDI schemes. Initial results favor the AIS approach to FDI due to its rather undemanding adaptation capabilities to new environments. The presence of the FDI scheme suggests benefits for the interaction between the pilot and the upset conditions by improving the accuracy of the identification of each particular failure and decreasing the detection delays. DEDICATION To my family for being always supportive in every decision I made and for helping me become who I am today. iii ACKNOWLEDGEMENTS First and foremost, I would like to thank Prof. Perhinschi and Prof. Napolitano for giving me the opportunity to further extend my college education. Specials thanks go to Prof. Perhinschi for letting me take part of this groundbreaking research giving me the best guidance one can expect from an Advisor. I would also like to thank Dr. Bojan Cukic for eagerly agreeing to form part of my committee and for providing positive feedback towards the successful completion of this work. My parents, Olga and Vicente, deserve a special thank for always letting me choose my own path and giving me an education one cannot learn from books; their generosity is out of imaginable limits. I would also like to express my sincere appreciation to my siblings Javier and Mariana for always believing in me even more than myself. I would also like to extend my gratitude to my friends, now spread throughout the world, for being the siblings one gets to choose to walk side by side through life, I am su...

Weak thruster fault detection for autonomous underwater vehicle based on artificial immune and signal pre-processing

“…Next, these extracted feature points are clustered and the maximum radius (r s , the maximum radius of each element of the self-set) is calculated based on clustering algorithm in Sanchez. 20 Based on the steps, the self-set is determined.…”

“…At first, a certain number of detectors with an initial radius are randomly distributed in non-self space (the other part non-belonging to self in the whole space). And then two optimization criteria are considered from Sanchez 20 to adjust the radius of the detectors, including no overlap among the detectors and self and the maximum covered areas in the non-self. In this process, some detectors would be deleted if the optimized radius is less than a certain value (given a priori).…”

“…, matching test needs to be conducted between C and D. According to Sanchez, 20 the matching rule is selected as the Euclidean distance between the test feature point and any of the detectors. If the Euclidean distance is less than the radius of the detector, then the test point will be treated as fault point.…”

“…As mentioned in section ''Anomaly detection based on AIS,'' in the process of generating self-set, this article uses clustering algorithm in Sanchez 20 to calculate the elements' radius of the self-set, and after calculation, the maximum radius is 0.15. The initial number of detectors is 5000 and their initial radius is 0.5.…”

“…This means that AUV velocity and attitude cannot directly reflect the thruster fault effect. Therefore, when directly adopting these signals as input of the AIS algorithm 20 for fault diagnosis, it is found that the detection result is not satisfactory, that is, the false alarm or undetected fault always occurs. Therefore, before conducting fault detection based on AIS, we need to extract the feature from the available AUV signals.…”