Research On Fault Tolerant Control And Event-Triggered Control For Uncertain Nonlinear Systems

Due to the increasing complexity of modern systems and the presence of model uncertainties,the higher demand on the safety and reliability of the system has been received much more attention.In practical engineering,unexpected faults of components including sensors and actuator and/or the system's structure occur inevitably,which may bring huge financial loss and social loss to the social society.It is thus essential and significant to develop effective fault-tolerant control designs to accommodate component failures,and maintain high degree safe operation performances of the overall control systems.On the other hand,with the rapid development of the network technique,the amount of data transmission in the network is becoming larger.Note that,based on the traditional time-triggered control method,the control signal will be periodically transferred and updated at the fix time intervals,which will cause a lot of unnecessary waste of redundant infor-mation to be transmitted.This is not desirable especially when the network resources are limited.For this reason,how to effectively reduce the amount of data transmission in the network,thereby saving bandwidth resources,has become an important research topic.Recently,the rapid development of the event-triggered control method has been paid more and more attention because of its unique aperiodic sampling characteristics.In recent years,a large number of research results have been emerged on the aforementioned two types of problems.However,due to the high degree of nonlinearity,uncertainty,multi-variable and strong coupling,it is difficult to meet the requirements of the control quality of the complex nonlinear system by using the traditional event-triggered control method and fault tolerant control technology.Therefore,there are still many problems need to be further studied.On the basis of the previous work,this dissertation further studies some problems of nonlinear uncertain systems,such as,adaptive event-triggered controller design,adaptive fault tolerant controller design and stability analysis of closeloop systems.The fault tolerant control is discussed for a class of parameter strict-feedback nonlinear systems with input quantization and actuator faults.Further,for a class of Lipschitz nonlinear systems with multiple delayed-state perturbations and actuator faults,an adaptive fault tolerant controller is designed without fault detection and fault diagnosis.For a class of strict-feedback nonlinear uncertain systems,a simultaneous design of a model-based adaptive neural control law and an adaptive event-triggering condition is investigated.Compared with the existing literatures,the restriction on the matching condition has been completely removed.For a class of pure-feedback nonlinear uncertain systems,a simultaneous design of a zero-order hold(ZOH)-based adaptive neural control law and an adaptive event-triggering condition is investigated.For a class of multi-input and multi-output strict-feedback nonlinear uncertain systems,a simultaneous design of a ZOH-based adaptive neural control law and an event-triggering condition is addressed.Parts of the developed theories are applied to the controller design of the continuous torsional pendulum system,m-link robot system,double-inverted pendulums model and the one single-link robot system.Simulation examples illustrate the advantages and effectiveness of proposed approaches.The main contents are outlined as follows:In Chapters 1-2,the development and main research methods of faulttolerant control and event-triggered control are analyzed and summarized,and some preliminaries about the considered problems are given.In Chapter 3,an adaptive fault-tolerant control problem is considered for a class of parameter strict-feedback nonlinear systems with input quantization,external disturbances and actuator faults.Firstly,an intermediate control law is designed by a modified adaptive backstepping design procedure,where a damping term with the estimate of unknown bounds and a positive time-varying integral function are introduced in the inter-mediate control law.Then,'a novel smooth function is introduced in the control law to eliminate the effect of quantization based on the intermediate control law constructed in the first step.Compared with existing literatures,the proposed scheme has the following advantages:1)the total number of failures is allowed to be infinite;2)the global stabil-ity of the overall closed-loop system is achieved,and the output tracking error converges to zero asymptotically in spite of input quantization,disturbances and possibly infinite number of faults.Finally,simulation results demonstrate the efficiency of the proposed algorithm.In Chapter 4,an adaptive fault-tolerant control problem is investigated for a class of Lipschitz nonlinear systems with multiple delayed-state perturbations and actuator faults including outage,loss of effectiveness and stuck.By introducing a positive nonlinear control gain function,a bound estimation approach and an integrable auxiliary signal,an adaptive fault-tolerant control scheme is proposed without fault detection mechanism.The effects of the unknown time-delay functions,external disturbances,and actuator faults can be compensated completely by the proposed control scheme.Furthermore,It is shown that the closed-loop signals are bounded and the system state converges to zero asymptotically despite unknown time delay function,actuator faults,and external disturbances.Simulation results are provided to illustrate the effectiveness of the proposed controllers.In Chapter 5,a model-based adaptive event-triggered controller design problem for a class of strict-feedback nonlinear uncertain systems is investigated.An adaptive model is firstly designed with event sampled state vector by using neural networks to approximate the nonlinear uncertainties.Based on the proposed adaptive model,a simultaneous design of a model-based adaptive neural control law and an adaptive event-triggering condition is investigated.A unified model for nonlinear impulsive systems including system states,model states and adaptive parameters is constructed.By using the Lyapunov function for impulsive dynamical systems,the stability of the closed-loop systems and the related convergence will be studied based on the adaptive event-triggering condition.Compared with the existing event-triggered literatures,the restrictions on the matching condition have been completely removed.Compared with traditional backstepping schemes with a continuous transmission,the proposed method only uses the sampling states to update the controller and the adaptive law,which can largely reduce the transmission load.Two examples are given to illustrate the effectiveness and merits of the proposed method.In Chapter 6,a ZOH-based adaptive event-triggered controller design problem for a class of pure-feedback nonlinear uncertain systems is investigated.The implicit function theorem and the mean value theorem are firstly used to transform the closed-loop system into a semiaffine form.A simultaneous design of a ZOH-based adaptive neural control law and an adaptive event-triggering condition is investigated by using backstepping technique.We formulate the event-triggered network control systems as nonlinear impulsive dynamical system,and a novel Lyapunov theorem is used to show the stability properties of the close-loop systems.Compared with traditional backstepping schemes with a continuous transmission,the proposed method can largely reduce the transmission load.Two examples are given to illustrate the effectiveness and merits of the proposed method.In Chapter 7,a ZOH-based adaptive event-triggered controller design problem for a class of multi-input and multi-output(MIMO)strict-feedback nonlinear uncertain systems is investigated.By using neural networks to approximate the nonlinear uncertainties,a simultaneous design of a ZOH-based adaptive neural control law and an adaptive event-triggering condition is investigated.Compared with the existing event-triggered literatures,the Lipschitz condition on the NN activation functions has been removed,thus the computation of NN weight estimates during interevent periods is not required.Two examples are given to illustrate the effectiveness and merits of the proposed method.Finally,the results of the dissertation are summarized and further research topics are pointed out.