Search

This paper focuses on adaptive tracking control for a class of strict-feedback nonlinear systems suffering from the replay attack in the sensor-controller channel. To address the challenges posed by repeating information and complex nonlinearity in nonlinear systems, an adaptive resilient controller based on the multidimensional Taylor network is designed. This controller guarantees that the tracking error converges to zero exponentially in the absence of replay attacks, and it achieves bounded tracking in the presence of replay attacks. Finally, the effectiveness of the proposed control strategy is verified using a chemical reactor with circulating flow simulation.

This paper focuses on the passive fault-tolerant control problem for a class of uncertain nonlinear discrete-time systems subject to multiple intermittent faults. The multiple intermittent faults considered occur on the actual measurement signals in the form of additive faults. To achieve the aim of fault-tolerant control, a dynamic output-feedback controller is designed such that, in the simultaneous presence of parameter uncertainty, a given type of nonlinearity, and multiple additive intermittent sensor faults, the closed-loop system remains stable and guarantees acceptable performance. The linear matrix inequality technique is used to obtain sufficient conditions to derive the controller and ensure that the prescribed H∞index is satisfied. Finally, the effectiveness of the proposed method is demonstrated though a numerical simulation example.

This study focuses on the individual sedation model identification problem, and proposes a subspace-based Wiener system identification method. The traditional compartmental pharmacokinetics-pharmacodynamics hypnosis model is considered as a specific Wiener system with the Hill equation nonlinear term. To deal with the Hill nonlinear term, the proposed method employs a set of specific bases to turn the Wiener system into a linear one, and then the subspace orthogonal projection identification method has been implemented to identify the transformed linear model. Compared with the traditional anesthesia model identification methods, the proposed method can effectively overcome the shortage of measurement data, get rid of the estimation of the effect compartment concentration which is impossible to be measured, and improve the individualization performance of the identified model. A simulation study on 24 various virtual patients from the Wang's Simulator has been conducted and validates the efficiency and robustness of the proposed method, and a drug infusion instruction has been provided in order to get relatively accurate identification performance.

Based on a two-dimensional (2D) system description of a batch process, a robust closed-loop iterative learning control (ILC) scheme is proposed for batch processes with time-varying uncertainties. An important merit is that the proposed ILC method can be used for on-line optimization against batch-to-batch process uncertainties to realize robust tracking of the setpoint trajectory in both the time (during a cycle) and batchwise (from cycle to cycle) directions. Only measured output errors of the current and previous cycles are used to design a synthetic ILC controller consisting of dynamic output feedback plus feedforward control. By introducing a comprehensive 2D difference Lyapunov function that can lead to monotonical state energy decrease in both the time and batchwise directions, sufficient conditions are established in terms of linear matrix inequality (LMI) constraints for holding robust stability of the closed-loop ILC system. By solving these LMI constraints, the ILC controller is explicitly formulated, together with an adjustable robust H infinity performance level. An illustrative example is given to demonstrate effectiveness and merits of the proposed ILC method.

A novel combination of PI control and iterative learning control (ILC), referred to ILC-based PI control, is proposed in this study. This algorithm belongs to set-point-related (SPR) indirect ILC. Based on a 2-dimensional (2D) Roesser's system description, a sufficient condition for asymptotical stability of multi-input multi-output linear batch processes under the ILC-based PI controller was derived in this paper. Applications on injection packing pressure control show that the proposed method can achieve the design objective well, with performance improvement along both time and batch direction, and also owns good robustness to disturbances and measurement noises.