具有输入输出约束特性的非线性系统自适应模糊控制

发布时间：2018-12-13 14:59

【摘要】：从控制工程的角度讲,系统的最终控制性能既和被控对象有关,又受到控制回路中如执行器和传感器等各种物理器件的性能以及通讯信道性能的影响。一方面,被控对象由于受到系统建模误差和工作环境等因素的影响通常具有本质非线性和不确定性。另一方面,执行器和传感器往往存在非光滑、非线性约束特性。当控制信号和输出信号经过这些约束环节时,就会引起系统性能恶化甚至出现系统失稳现象。此外,通讯信道受到网络带宽的限制,控制信号在传输之前进行量化在所难免,由此产生的量化误差同样会对系统控制性能造成极大的负面影响。鉴于此,本文以backstepping技术为框架,以模糊逻辑系统作为函数逼近器,系统地研究具有输入输出约束特性的非线性系统的自适应控制问题。本文共分六章。第一章概述具有输入输出约束的非线性系统研究现状。从第二章开始,主要研究内容从五个部分展开,每一部分作为一章。第二章针对具有未建模动态和动态干扰的非线性系统,提出了一种直接自适应模糊输出反馈控制方案。在设计过程中,引入一个线性状态观测器来估计系统状态,采用模糊逻辑系统(Fuzzy Logic Systems, FLS)逼近未知的虚拟控制信号,结合反步递推方法设计了一种自适应模糊控制器,借助小增益定理证明了闭环系统的输入-状态稳定性。该控制方案放宽了以往文献中对动态干扰项的强假设条件,并通过在线估计模糊逻辑系统权向量的范数,减少了在线调节的自适应参数个数,从而加快了自适应控制算法的在线运行效率。第三章充分考量复杂工作环境下执行器死区的不确定性和摄动特性,创新性地提出了一种模糊死区模型,以研究具有不确定死区输入的非线性系统的跟踪控制问题。结合模糊集理论和集成控制思想,首先,针对具有不可测量状态和模糊死区输入的严格反馈非线性系统,提出了一种自适应综合控制设计方案。该方案保证了闭环系统的稳定性和跟踪性能。接着,研究了具有模糊死区输入的未建模动态非线性系统的跟踪控制问题,利用辅助动态信号控制未建模动态,结合集成控制思想和动态表面(DSC)技术,设计了新颖的自适应控制器。第四章针对实际控制系统中执行器磁滞方向容易发生跳变的情况,提出了变方向的Bouc-wen磁滞模型。基于该磁滞模型,研究了具有磁滞输入的随机纯反馈非线性系统的自适应跟踪控制问题。在引理5.1的基础上,通过引入一个辅助虚拟控制器并利用Nussbaum函数的性质,在随机非线性系统中解决了磁滞输入变方向的难题,结合backstepping技术,提出了一种新颖的自适应模糊控制设计方案。与已有的磁滞输入问题的研究工作相比,本章所考虑的系统更具一般性,从而扩展了磁滞输入问题的应用范围。第五章为抵消输出传动装置中的非线性环节对系统性能的负面影响,研究了具有未知输出死区的严格反馈非线性系统的跟踪控制问题。一方面,现有的输出非线性研究工作都集中在线性系统或者满足匹配条件的非线性系统的镇定问题上,其方法难以控制比较复杂的非线性系统(如严格反馈非线性系统)的跟踪控制问题。另一方面,实际系统中的状态变量常常难以获得,这导致以往包含了部分或全部状态变量的backstepping设计方案也不能直接用来控制该类系统。本文通过建立状态的非线性函数与输出之间的关系,引入一个Nussbaum函数和辅助虚拟控制器,提出了一种全新的控制器设计方法,解决了这类复杂系统的跟踪控制问题。第六章考虑到量化反馈控制在数字控制、网络化控制系统等领域中的广泛应用,研究了具有输入量化约束的随机非线性系统的性能控制问题。首先,利用磁滞类量化器的扇形有界性质提出了量化器输出的一种新的非线性分解策略,该策略克服了以往线性分解策略中扰动项的界不好确定的问题。接着,运用这种非线性分解策略,提出了一种新的自适应模糊控制方案,解决了具有输入量化的随机严格反馈非线性系统的跟踪控制问题。该方案通过在线学习机制补偿了量化误差,不需要系统和量化器参数满足强的假设条件,从而在有限通讯频率下仍能保证系统的跟踪性能。然后,充分考量未建模动态对量化反馈非线性系统的负面影响,研究了具有量化输入约束的未建模动态随机非线性系统的镇定问题。结合反步递推技术和小增益方法,提出了全新的自适应模糊控制方案,保证了闭环系统是依概率输入-状态稳定的。
[Abstract]:From the angle of control engineering, the final control performance of the system is not only related to the controlled object but also the performance of various physical devices such as the actuator and the sensor in the control loop, as well as the effect of the communication channel performance. On the one hand, the controlled object is generally nonlinear and uncertain due to the influence of factors such as system modeling error and working environment. on the other hand, the actuators and sensors often have non-smooth, non-linear constraint characteristics. When the control signal and the output signal pass through these restriction links, the system performance degradation and even the system instability can be caused. In addition, the communication channel is limited by the network bandwidth, the control signal is quantified before transmission, and the resulting quantization error also has a great negative effect on the system control performance. In view of this, the adaptive control of nonlinear systems with input and output constraint characteristics is systematically studied by using the backstepping technique as a frame and using the fuzzy logic system as a function approximation. This article is divided into six chapters. The first chapter provides an overview of the research status of nonlinear systems with input and output constraints. From the second chapter, the main research contents are expanded from five parts, each part as a chapter. In the second chapter, a direct adaptive fuzzy output feedback control scheme is proposed for nonlinear systems with unmodeled dynamic and dynamic interference. In the design process, a linear state observer is introduced to estimate the state of the system. The fuzzy logic system (FLS) is used to approximate the unknown virtual control signal, and a self-adaptive fuzzy controller is designed in combination with the reverse-step recursion method. The input-state stability of closed-loop system is proved by means of small gain theorem. The control scheme has the advantages that the strong hypothesis condition of the dynamic interference term in the prior art is relaxed, the norm of the weight vector of the fuzzy logic system is estimated on-line, the number of the self-adaptive parameters of the on-line adjustment is reduced, and the on-line running efficiency of the adaptive control algorithm is accelerated. In the third chapter, the uncertainty and the perturbation characteristic of the dead zone of the actuator in the complex working environment are fully considered, and a fuzzy dead zone model is proposed to study the tracking control problem of the nonlinear system with uncertain dead zone input. Combined with the theory of fuzzy set and the idea of integrated control, a self-adaptive comprehensive control scheme is proposed for the strict feedback nonlinear system with non-measurable state and fuzzy dead zone input. The scheme guarantees the stability and tracking performance of the closed-loop system. Then, the tracking control problem of the unmodeled dynamic nonlinear system with fuzzy dead zone input is studied, and the new adaptive controller is designed by using the auxiliary dynamic signal to control the unmodeled dynamics, combining the integrated control idea and the dynamic surface (DSC) technology. In the fourth chapter, the change of the hysteresis of the actuator in the actual control system is easy to jump, and the Bouc-wen hysteresis model of the variable direction is put forward. Based on the hysteresis model, the self-adaptive tracking control problem of a stochastic pure-feedback nonlinear system with hysteresis input is studied. On the basis of lemma, a novel adaptive fuzzy control scheme is proposed by introducing an auxiliary virtual controller and using the properties of the Nusculum function. In the random nonlinear system, a novel adaptive fuzzy control design scheme is proposed. Compared with the existing research work of the hysteresis input problem, the system considered in this chapter is more general, thus extending the application range of the hysteresis input problem. The fifth chapter is to cancel the negative influence of the non-linear link in the output transmission device on the system performance, and to study the tracking control problem of the strict feedback nonlinear system with unknown output dead zone. On the one hand, the existing output non-linear research work is focused on the stabilization problem of a linear system or a non-linear system satisfying the matching condition, and the method is difficult to control the tracking control problem of a complex nonlinear system (such as a strict feedback nonlinear system). On the other hand, the state variables in the actual system are often difficult to obtain, which results in a backstepping design that has previously included some or all of the state variables and cannot be used directly to control such systems. In this paper, a new controller design method is proposed to solve the tracking control problem of this kind of complex system by establishing the relation between the non-linear function and the output of the state, introducing a Nusculum function and an auxiliary virtual controller. The sixth chapter, taking into account the wide application of the quantitative feedback control in the fields of digital control and networked control system, has studied the performance control problem of the stochastic nonlinear system with input quantization constraint. First, a new non-linear decomposition strategy for the output of a quantizer is proposed by using the sector-specific property of the hysteresis class quantizer, which overcomes the problem that the boundary of the perturbation term in the prior linear decomposition strategy is not well defined. Then, using this non-linear decomposition strategy, a new adaptive fuzzy control scheme is proposed to solve the problem of tracking control with input quantization and random strict feedback nonlinear system. The scheme can compensate the quantization error through the on-line learning mechanism, and does not need the system and the quantizer parameter to meet the strong hypothesis condition, so that the tracking performance of the system can be ensured under the limited communication frequency. Then, the negative influence of the unmodeled dynamics on the quantitative feedback nonlinear system is fully considered, and the stabilization problem of the unmodeled dynamic random nonlinear system with the quantized input constraint is studied. In this paper, a new self-adaptive fuzzy control scheme is proposed in combination with the reverse-step recursive technique and the small-gain method, which ensures that the closed-loop system is stable according to the probability input-state.
【学位授予单位】：广东工业大学
【学位级别】：博士
【学位授予年份】：2015
【分类号】：TP273.4

【相似文献】