In this paper, two approaches are proposed and compared for the detection and identification of aircraft subsystem failures based on the artificial immune system paradigm combined with the hierarchical multiself strategy. The first approach relies on the heuristic ranking of lower order self/non-self projections and the generation of selective immunity identifiers through structuring of the non-self. The second approach is based on an information processing algorithm inspired by the functionality of the dendritic cells. The artificial dendritic cell is defined as a computational unit that centralizes, fuses, and interprets information from the multiple selves to produce a unique detection and identification outcome. A hierarchical multi-self strategy is used with both approaches considering 2-dimensional self/non-self projections or subselves. A mathematical formulation of the concepts and detailed implementation algorithms are presented. The proposed methodologies are demonstrated and compared using simulation data for a supersonic fighter from a motion-based flight simulator at nominal conditions, under failures of actuators, malfunction of sensors, and wing damage. In all cases considered, both detection and identification schemes achieve excellent detection and identification rates with practically no false alarms.
Immunity-Based Accommodation of Aircraft Subsystem Failures This thesis presents the design, development, and flight-simulation testing of an artificial immune system (AIS) based approach for accommodation of different aircraft subsystem failures. Failure accommodation is considered as part of a complex integrated AIS scheme that contains four major components: failure detection, identification, evaluation, and accommodation. The accommodation part consists of providing compensatory commands to the aircraft under specific abnormal conditions based on previous experience. In this research effort, the possibility of building an AIS allowing the extraction of pilot commands is investigated. The proposed approach is based on structuring the self (nominal conditions) and the non-self (abnormal conditions) within the AIS paradigm, as sets of artificial memory cells (mimicking behavior of T-cells, B-cells, and antibodies) consisting of measurement strings, over pre-defined time windows. Each string is a set of features values at each sample time of the flight including pilot inputs, system states, and other variables. The accommodation algorithm relies on identifying the memory cell that is the most similar to the incoming measurements. Once the best match is found, control commands corresponding to this match will be extracted from the memory and used for control purposes. The proposed methodology is illustrated through simulation of simple maneuvers at nominal flight conditions, different actuators, and sensor failure conditions. Data for development and demonstration have been collected from West Virginia University 6-degreesof-freedom motion-based flight simulator. The aircraft model used for this research represents a supersonic fighter which includes model following adaptive control laws based on non-linear dynamic inversion and artificial neural network augmentation. The simulation results demonstrate the possibility of extracting pilot compensatory commands from the self/non-self structure and the capability of the AIS paradigm to address the problem of accommodating actuator and sensor malfunctions as a part of a comprehensive and integrated framework along with abnormal condition detection, identification, and evaluation. DEDICATION To my mother Gulnara Yezhebayeva and my grandparents,
This paper presents a novel bio-inspired adaptive control technique that has been designed to maintain the performance of an aircraft under upset conditions. The proposed control approach is inspired by biological principles that govern the humoral response of the immune system of living organisms and is intended to reduce pilot effort while maintaining adequate aircraft operation outside bounds of nominal design. The immunity-based control parameters are optimized offline for multiple sets of failures using a genetic algorithm approach. The performance of the immunity-based augmentation is compared with a neural network (NN)-based augmentation. Different piloted tests were performed on a six degrees-of-freedom (6DOF) motion-based simulator for different types of maneuvers under several flight conditions. The results show that the artificial immune system (AIS) proposed scheme improves the aircraft handling qualities by reducing the tracking errors (TEs) and improving the pilot response required to maintain control of the aircraft under upset conditions.
This paper presents the development and testing of a novel fault tolerant adaptive control system based on a bio-inspired immunity-based mechanism applied to an aircraft fighter model. The proposed baseline control laws use a non-linear dynamic inversion and model reference adaptive control on the inner loops of the aircraft dynamics. In this new approach, the baseline controllers are augmented with an artificial immune system mechanism that relies on a direct compensation inspired primarily by the biological immune system response. The effectiveness of the approach is demonstrated through a full 6 degrees-of-freedom aircraft model interfaced with a Flight gear environment. The performance of the proposed control laws are investigated under a novel set of performance metrics, which quantify the level of input activity from the pilot and from the control surfaces in order to ensure the stability and performance of the aircraft under different actuator and structural failures. Optimization of the parameters of the artificial immunity system is performed using a genetic algorithm. The results show that the optimized fault tolerant adaptive control laws improve significantly the failure rejection using minimum pilot input and control surfaces activity under upset flight conditions.
This paper presents the development of a biologically-inspired methodology for flight envelope prediction at post failure conditions. The flight envelope is understood in its most general meaning as the hyper-space of all achievable or desirable relevant variables. The new ranges of these variables at post-failure conditions are the outcomes of the prediction process. Specific algorithms are proposed depending on the affected subsystem and the nature and characteristics of the failure. Actuator, sensor, propulsion system, and structural failures are considered. The proposed methodology is integrated with immunity-based failure detection and identification and benefits from the capabilities of the artificial immune system to address directly the complexity and multidimensionality of aircraft dynamic response in the context of abnormal conditions. A hierarchical multi-self strategy is used, in which low-dimensional projections replace the hyperspace of the self thus avoiding numerical and conceptual issues related to the high-dimensionality of the problem. The methodology is illustrated through numerical examples of envelope prediction under elevator locked failure, yaw rate sensor bias, locked throttle, and partially missing horizontal tail.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.