Prescribed performance adaptive neural output control for a class of switched nonstrict‐feedback nonlinear time‐delay systems: State‐dependent switching law approach

This paper investigates the tracking problem for a class of uncertain switched nonlinear delayed systems with nonstrict‐feedback form. To address this problem, by introducing a new common Lyapunov function (CLF), an adaptive neural network dynamic surface control is proposed. The state‐dependent switching rule is designed to orchestrate which subsystem is active at each time instance. In order to compensate unknown delay terms, an appropriate Lyapunov‐Krasovskii functional is considered in the constructing of the CLF. In addition, a novel switched neural network–based observer is constructed to estimate system states through the output signal. To maintain the tracking error performance within a predefined bound, a prescribed performance bound approach is employed. It is proved that by the proposed output‐feedback control, all the signals of the closed‐loop system are bounded under the switching law. Moreover, the transient and steady‐state tracking performance is guaranteed by the prescribed performance bound. Finally, the effectiveness of the proposed method is illustrated by two numerical and practical examples.     

New observer‐based controller design for LPV stochastic systems with multiplicative noise

This paper addresses an observer‐based control problem of Linear Parameter Varying (LPV) stochastic systems. Based on the modeling approaches, the LPV stochastic systems can be represented by a set of linear systems with multiplicative noise term. To solve the observer‐based control problem, a less conservative stability criterion is developed via the chosen Parameter‐Dependent Lyapunov Function (PDLF). In the PDLF, none element in the positive definite matrix is required as zero. Besides, an Extended Projection Lemma is proposed to convert the derived sufficient conditions into Linear Matrix Inequality (LMI) form. According to the derived LMI conditions, all feasible solutions can be found by convex optimization algorithm at a single step. Based on those feasible solutions, an observer‐based Gain‐Scheduled (GS) controller can be established to guarantee the asymptotical stability of the closed‐loop system in the sense of mean square. Finally, two numerical examples are provided to demonstrate the effectiveness and applicability of the proposed method.     

Observer‐based mode‐independent integral sliding mode controller design for phase‐type semi‐Markov jump singular systems

This paper considers the observer‐based integral sliding mode controller design problem of semi‐Markovian jumping singular systems with time‐varying delays. Firstly, by using a plant transformation and supplementary variable technique in the work of Hou et al, the discussed phase‐type semi‐Markov jump singular system is equivalently transformed into its associated Markov jump singular system. Secondly, an observer‐based sliding mode controller design problem is investigated for the associated singular Markov jump systems. The highlight of this paper is that we construct an observer‐based mode‐independent integral sliding mode surface function, which is different from the mode‐dependant sliding mode surface function in the previous literatures. Based on this, an observer‐based sliding mode controller is designed to guarantee that the associated singular Markov jump system meets the reachable condition. Finally, a practical example is presented to demonstrate the efficiency and effectiveness of our obtained results.     

Adaptive diagnosis and compensation for hypersonic flight vehicle with multisensor faults

This study proposes an improved adaptive fault diagnosis and compensation scheme for multisensor faults of hypersonic flight vehicles (HFVs). The faults are detected and isolated through a series of sensor output residuals and thresholds that consider observation error and disturbances. Via an adaptive augmented observer, the faults are estimated accurately and a time‐varying disturbance is handled by an additional differential part. Sensor faults are compensated on the basis of estimation results, and disturbances are considered in the fault‐tolerant control (FTC) design, thereby improving the tracking accuracy of the altitude and velocity and robustness with respect to external disturbances for HFVs. The stability of diagnosis and FTC is analyzed by Lyapunov theory. Numerical simulation results explain the validity of the proposed diagnosis and compensation methods.     

自适应广义滑模观测器之状态估计和故障重构

针对一类同时具有执行机构故障和输出扰动的不确定性系统,提出了一种自适应广义滑模观测器,实现系统状态的估计和执行机构故障的重构.首先,进行系统变换及状态增广构成一类奇异系统,使得执行机构故障和输出扰动解耦,且输出扰动成为增广系统的状态之一;随后,受未知输入观测器和滑模观测器的启发,设计了一种广义滑模观测器,放宽了传统方法对输出维数的严格约束,并通过融入自适应技术放宽了对故障和不确定性上界已知的要求;最后,利用等效误差注入的思想实现故障重构.针对可重复使用运载器再入段进行的仿真实例,验证了所提方法的有效性.     

油藏注水开发低阶模型及最优控制仿真研究

油藏注水开发最优控制问题计算规模大、控制变量与计算网格多,且控制变量与目标函数之间的关系为一组非线性偏微分方程控制,若直接进行数值求解,对于目前的计算机计算速度和存储空间是个巨大负担.本文采用最佳正交分解(proper orthogonal decomposition,POD)方法提出了基于低阶模型的油藏注水开发最优控制问题,这样,控制变量与目标函数之间的复杂关系被转变为解析函数,仅以少量的POD系数作为优化变量且只需采用非线性规划方法即可求解,大幅度地降低了原问题的求解复杂度与计算量.以二维五点井网的一个井组为应用实例进行仿真研究,结果表明:基于低阶模型的最优控制问题所求解的最大生产净现值与经典的伴随梯度法相比仅有不超过2.5%的误差,且计算速度优势极为明显,当网格数为40×40时,计算速度可提高30倍以上,网格数越多,计算速度优势越明显,当网格数为70×70时,可提速60倍以上.     

可逆冷带轧机速度张力系统的模糊自适应动态面反步控制

研究了基于动态面反步控制和模糊自适应逼近的可逆冷带轧机速度张力系统直接反馈线性化(direct feedback linearization,DFL)动态解耦控制问题.首先,通过构造非线性干扰观测器(nonlinear disturbance observer,NDO)削弱了模型中非匹配不确定项对系统性能的影响,进而应用DFL理论实现了速度张力非线性耦合系统的动态解耦和线性化;其次,将反步控制与动态面控制相结合完成了解耦后速度张力各线性子系统控制器的设计,且有效避免了反步控制中的"微分爆炸"现象;再次,采用模糊自适应方法对所设计控制器中的匹配不确定项进行了逼近估计,有效地提高了速度张力系统的跟踪控制精度;稳定性分析结果表明,可逆冷带轧机速度张力系统是一致最终有界的.最后,基于工业现场的实际数据进行仿真对比研究,仿真结果验证了本文所提方法的有效性.     

航天器姿态输出反馈抗干扰跟踪控制

针对采用四元数描述的航天器姿态运动模型研究输出反馈抗干扰跟踪控制问题.首先在姿态运动模型的基础上,结合四元数的性质设计扩张状态观测器(extended state observer,ESO)来估计角速度和干扰力矩,从理论上保证了ESO中的四元数状态满足范数约束,并证明了观测误差的收敛性;进一步利用互连和阻尼分配无源控制(interconnection and damping assignment passivity-based control,IDA--PBC)理论设计控制律,通过姿态和角速度误差状态变换以及引入误差积分项,使得期望的姿态和角速度误差,以及积分项误差运动方程中均出现阻尼项,提高了系统的抗干扰性能,最后利用Laypunov函数证明了闭环系统一致最终有界稳定.仿真结果验证了所设计ESO和IDA--PBC控制律的有效性.     

弱滑模观测器在无刷直流电机中的仿真研究

深入讨论了基于幂次趋近率方法的滑模观测器,并对一系列切换函数进行对比研究,总结出弱滑模观测器的规律.通过软硬件方式间接获得可靠的转子位置信号是无位置传感器无刷直流电机矢量控制的基础.在状态观测器设计法中,基于幂次趋近率方法的滑模观测器对电机启动时的反电动势信号具有跟踪速度快、观测精度高等特点,且算法结构简单,便于实现.仿真结果表明该方法在电机启动时反电动势信号微弱的情况下可以准确地估算反电动势信号,快速跟踪实际信号,并证明了该方法的正确性与有效性.     

模糊滑模控制和负载转矩补偿的异步电动机直接转矩控制

针对异步电动机(IM)转矩脉动以及抗干扰能力差的问题,设计了基于模糊滑模控制(FSMC)与负载转矩补偿的新型直接转矩控制(DTC),取代传统PID速度调节器的是一种滑模控制器.为解决滑模控制器中负载转矩脉动的问题,用模糊逻辑控制器取代了传统滑模控制律中的不连续部分,可以明显降低异步电动机在低速运转时的转矩脉动.提出了一种负载转矩观测器来估计未知的负载转矩.负载转矩观测器用来估计负载转矩扰动,估计作为速度环的前馈补偿.仿真结果表明:在低速负载转矩扰动时,该设计具有更好的动态响应和速度性能、更高鲁棒性和更强的抗干扰能力.