Integrated Immunity-based Methodology for UAV Monitoring and Control

A general integrated and comprehensive health management framework based on the artificial immune system (AIS) paradigm is formulated and an automated system is developed and tested through simulation for the detection, identification, evaluation, and accommodation (DIEA) of abnormal conditions (ACs) on an unmanned aerial vehicle (UAV). The proposed methodology involves the establishment of a body of data to represent the function of the vehicle under nominal conditions, called the self, and differentiating this operation from that of the vehicle under an abnormal condition, referred to as the non-self. Data collected from simulations of the selected UAV autonomously flying a set of prescribed trajectories were used to develop and test novel schemes that are capable of addressing the AC-DIEA of sensor and actuator faults on a UAV. While the specific dynamic system used here is a UAV, the proposed framework and methodology is general enough to be adapted and applied to any complex dynamic system. The ACs considered within this effort included aerodynamic control surface locks and damage and angular rate sensor biases. The general framework for the comprehensive health management system comprises a novel complete integration of the AC-DIEA process with focus on the transition between the four different phases. The hierarchical multiself (HMS) strategy is used in conjunction with several biomimetic mechanisms to address the various steps in each phase. The partition of the universe approach is used as the basis of the AIS generation and the binary detection phase. The HMS approach is augmented by a mechanism inspired by the antigen presenting cells of the adaptive immune system for performing AC identification. The evaluation and accommodation phases are the most challenging phases of the AC-DIEA process due to the complexity and diversity of the ACs and the multidimensionality of the AIS. Therefore, the evaluation phase is divided into three separate steps: the qualitative evaluation, direct quantitative evaluation, and the indirect quantitative evaluation, where the type, severity, and effects of the AC are determined, respectively. The integration of the accommodation phase is based on a modular process, namely the strategic decision making, tactical decision marking, and execution modules. These modules are designed by the testing of several approaches for integrating the accommodation phase, which are specialized based on the type of AC being addressed. These approaches include redefining of the mission, adjustment or shifting of the control laws, or adjusting the sensor outputs. Adjustments of the mission include redefining of the trajectory to remove maneuvers which are no longer possible, while adjusting of the control laws includes modifying gains involved in determination of commanded control surface deflections. Analysis of the transition between phases includes a discussion of results for integrated example cases where the proposed AC-DIEA process is applied. The cases considered show the validity of the integrated AC-DIEA system and specific accommodation approaches by an improvement in flight performance through metrics that capture trajectory tracking errors and control activity differences between nominal, abnormal, and accommodated cases