This paper presents 2-novel linear matrix inequality (LMI)-based adaptive output feedback fault-tolerant control strategies for the class of nonlinear Lipschitz systems in the presence of bounded matched or mismatched disturbances and simultaneous occurrence of actuator faults, including failure, loss of effectiveness, and stuck. The constructive algorithms based on LMI with creatively using Lyapunov stability theory and without the need for an explicit information about mode of actuator faults or fault detection and isolation mechanism are developed for online tuning of adaptive and fixed output-feedback gains to stabilize the closed-loop control system asymptotically. The proposed controllers guarantee to compensate actuator faults effects and to attenuate disturbance effects. The resulting control methods have simpler structure, as compared with most existing recent methods and more suitable for practical systems. The merits of the proposed fault-tolerant control scheme have been verified by the simulation on nonlinear Boeing 747 lateral motion dynamic model subjected to actuator faults. KEYWORDS actuator fault, fault-tolerant control (FTC), linear matrix inequality (LMI), Lipschitz nonlinear systems, output feedback, robust adaptive control
INTRODUCTIONA permanent increase in the complexity, efficiency, and reliability of nonlinear systems from both the theoretical and practical aspects necessitates a continuous development in fault-tolerant control (FTC) theory and practice. In this way, the considerable interest in FTC system design has received great attention over the past decades. Fault-tolerant control is a control technique that provides the ability to maintain overall linear and nonlinear system stability besides the acceptable performance in the event of component (including sensors, actuators, and even the plant itself) faults. The faults may occur in each of the system components at an uncertain time, and the size of the faults is also unknown. 1 This means that for the design of the controller for each system, the probability of fault occurrences in system components should be considered, and then, the controller design must provide a way to compensate them.The latest results and published papers confirm that there are still some challenging areas within the FTC for nonlinear systems on methodologies and computational complexities. For instance, most FTC methods only considered some of the component faults or the design methods are dependent on the states of systems, 2-5 and then, implementation of these methods for a large domain of real applications will be hardly possible.