International Journal of Robust and Nonlinear Control最新文献

The article focuses on the approximation of large-scale nonlinear systems by families of nearly independent systems, subject to a specified decentralization error bound. This creates an approximation of the original large-scale system by a decentralized system, where the deviation between the original system and the decentralized system does not exceed a specified error bound. The approximation is then used to design decentralized optimal robust controllers that control the original large-scale nonlinear system. The optimization objective is tracking, that is, maintaining the system's response as close as possible to a specified target state. Implementable controllers that approximate optimal performance are presented.

Modern control systems are expected not only to pursue optimality, but also to dynamically adapt to varying environments. Bridging the gap between adaptive control, optimal control, and data-driven learning control, reinforcement learning has emerged as a computationally efficient approach to achieve adaptive optimal control of systems. This paper surveys both theoretical advancements and practical applications of control oriented reinforcement learning methods, especially adaptive dynamic programming (ADP). We discuss recent progress of ADP in several key control disciplines, including optimal control, robust control, event-triggered control, distributed control and safe control; as well as real-world applications of ADP in various scenarios such as unmanned vehicles, power systems, intelligent transportation, robot manipulators, and motors.

This article focuses on the detection of false data injection (FDI) attacks in nonlinear cyber-physical systems with uncertain parameters, aiming to improve both the speed and accuracy of actuator detection affected by FDI attacks. First, considering the uncertainties or errors often present in practical cyber-physical system models, two $��{�\mathscr{H}�}_{�\infty �}�$ FDI attack detection estimator algorithms with different levels of conservatism are proposed to enhance detection accuracy. Second, compared to traditional detection thresholds, a novel adaptive detection threshold is introduced to improve the speed of attack detection. Third, to address the issue of limited communication bandwidth, transmission signals are quantized. Finally, the applicability and effectiveness of the proposed method are validated through two simulation examples.

This paper studies a model-free adaptive iterative learning heading control problem for unmanned surface vehicles (USVs) with quantized data and data dropouts. First, a compact form dynamic linearization model is established for USVs using a dynamic linearization technique and a redefined scheme. To address data dropout issues, a comprehensive compensation strategy is developed. In addition, a logarithmic quantization mechanism is introduced to reduce the transmission burden. Based on these elements, a quantized model-free adaptive iterative learning control approach is designed. The convergence of the heading control error of USVs governed by the proposed method is rigorously proven, and its effectiveness is verified through simulation results.

This article addresses the stabilization problem of networked control systems (NCSs) under non-periodic denial-of-service (DoS) attacks and actuator saturation. A non-periodic DoS attack model is put forward, taking into account the minimum communication security time period, the maximum attack duration, and the frequency constraint. This model facilitates the analysis of system performance in relation to attack characteristics by establishing connections between two generalized assumption models. For the NCSs suffering from both non-periodic DoS attacks and actuator saturation, a dynamic memory-based event-triggered mechanism (DMETM) is designed to maintain control performance and reduce network resource consumption. To deal with the actuator saturation phenomenon, a set of memory-based primary-auxiliary controllers is designed to compensate for actuator saturation effects via convex hull representation. A guaranteed-cost function accounting for system states and input saturation is developed. Based on the obtained switched system model and the piecewise Lyapunov-Krasovskii functional (LKF), sufficient conditions are derived to ensure local asymptotic stability and weighted disturbance attenuation with $��{�H�}_{�\infty �}�$ performance. The co-design of DMETM weighting matrices and controller gains is presented in the form of linear matrix inequalities (LMIs). Additionally, methods for estimating the attraction domain and deriving the upper bound of the cost function are given by the obtained LMIs. Finally, the effectiveness and superiority of the proposed method are validated on a high-speed train system.

In this paper, a cooperative guidance law is put forward for the multi-leader-multi-follower system with impact angle constraint. First, the multi-leader guidance without central coordination is described as a Nash game problem. Under the unwanted influence caused by guidance uncertainties, we use a certain upper bound of the Nash cost function to replace the original one. Then, the so-called near-optimal Nash guidance law for the multi-leader is obtained by using the Riccati-like differential equation. Second, to achieve a salvo attack by the followers in the same group of one leader, a guidance law for the follower along the line-of-sight (LOS) is provided by using the appointed-time sliding mode control method. Third, in the normal LOS direction of the follower, a finite-time super-twisting sliding mode guidance law is designed to ensure that the follower converge to the desired impact angle within a finite time. In the end, numerical simulations validate the effectiveness of the proposed guidance laws.

This article investigates the optimal control problem with disturbance rejection for discrete-time multi-agent systems under cooperative and non-cooperative graphical games frameworks. Given the practical challenges of obtaining accurate models, Q-function-based policy iteration methods are proposed to seek the Nash equilibrium solution for the cooperative graphical game and the distributed minmax solution for the non-cooperative graphical game. To implement these methods online, two reinforcement learning frameworks are developed, an actor-disturber-critic structure for the cooperative graphical game and an actor-adversary-disturber-critic structure for the non-cooperative graphical game. The stability of the proposed methods is rigorously analyzed, and simulation results are provided to illustrate the effectiveness of the proposed methods.