Annual Reviews in Control最新文献

In recent years, cybersecurity has received increasing attention due to the demand from a large class of networked systems for resilience against cyberattacks that may compromise privacy, integrity and availability. Many of these systems are abstracted as Discrete Event Systems (DES) as their evolution occurs through the occurrence of discrete events. Since they use communication networks and consequently may be vulnerable to attacks, cybersecurity must be considered in DES. Based on this challenging scenario, this work focuses on cybersecurity strategies in the context of DES. A systematic literature mapping (SLM) was carried out, which selected 208 papers on the aforementioned topic. These papers were analyzed and categorized regarding the characteristics of each cybersecurity strategy, the types of attacks considered and also the modeling formalism used. The primary objective of this work is to collect all relevant research in the literature to provide the state of the art on cybersecurity strategies for DES, as well as identify research trends and directions for future work on the topic. The results show that the majority of the selected papers present cybersecurity methods based on the strategy of protecting systems against passive attacks, using automata as the modeling formalism. In contrast, the topic of active attacks has gained attention in recent years, with an increasing number of papers published in several journals and conferences. Finally, research gaps and challenges are presented to provide future directions in the domain of the cybersecurity of DES.

Composite adaptation and learning techniques were initially proposed for improving parameter convergence in adaptive control and have generated considerable research interest in the last three decades, inspiring numerous robot control applications. The key idea is that more sources of parametric information are applied to drive parameter estimates aside from trajectory tracking errors. Both composite adaptation and learning can ensure superior stability and performance. However, composite learning possesses a unique feature in that online data memory is fully exploited to extract parametric information such that parameter convergence can be achieved without a stringent condition termed persistent excitation. In this article, we provide the first systematic and comprehensive survey of prevalent composite adaptation and learning approaches for robot control, especially focusing on exponential parameter convergence. Composite adaptation is classified into regressor-filtering composite adaptation and error-filtering composite adaptation, and composite learning is classified into discrete-data regressor extension and continuous-data regressor extension. For the sake of clear presentation and better understanding, a general class of robotic systems is applied as a unifying framework to show the motivation, synthesis, and characteristics of each parameter estimation method for adaptive robot control. The strengths and deficiencies of all these methods are also discussed sufficiently. We have concluded by suggesting possible directions for future research in this area.

When controlling cyber–physical systems via consensus algorithms, the robustness issue is of paramount importance because of model mismatching and/or disturbances that generally modify the Laplacian flow dynamics associated to the overall network, compromising the possibility to achieve any expected, orchestrated emergent behavior. To face this issue, the Variable Structure Control (VSC) approach has recently been shown to be an effective tool in designing control protocols over networks, providing robustness and often yielding finite-time convergence. Moreover, VSC further enables the decoupling of simultaneous control objectives where reaching a consensus state is only part of a more complex task, for instance as it happens in distributed optimization. Thus motivated, this paper overviews, from a tutorial perspective, some selected recent advances in the application of VSC with Sliding Modes to design robust consensus controllers in both the leader-less and the leader-following settings. Efforts are also made to unify the notation and to discuss the theoretical foundations of the nonsmooth analysis tools at the basis of their design for a wide readership unfamiliar with these problems and formal tools. To this aim, examples supporting the treatment, and numerical simulations, are given and discussed in detail. Finally, hints for future investigations along with some current open problems are provided.

This survey presents recent research on determining control-theoretic properties and designing controllers with rigorous guarantees using semidefinite programming and for nonlinear systems for which no mathematical models but measured trajectories are available. Data-driven control techniques have been developed to circumvent a time-consuming modelling by first principles and because of the increasing availability of data. Recently, this research field has gained increased attention by the application of Willems’ fundamental lemma, which provides a fertile ground for the development of data-driven control schemes with guarantees for linear time-invariant systems. While the fundamental lemma can be generalized to further system classes, there does not exist a comparable data-based system representation for nonlinear systems. At the same time, nonlinear systems constitute the majority of practical systems. Moreover, they include additional challenges such as data-based surrogate models that prevent system analysis and controller design by convex optimization. Therefore, a variety of data-driven control approaches has been developed with different required prior insights into the system to ensure a guaranteed inference. In this survey, we will discuss developments in the context of data-driven control for nonlinear systems. In particular, we will focus on methods based on system representations providing guarantees from finite data, while the analysis and the controller design boil down to convex optimization problems given as semidefinite programming. Thus, these approaches achieve reasonable advances compared to the state-of-the-art system analysis and controller design by models from system identification. Specifically, the paper covers system representations based on extensions of Willems’ fundamental lemma, set membership, kernel techniques, the Koopman operator, and feedback linearization.

Data-driven predictive control (DPC) has gained an increased interest as an alternative to model predictive control in recent years, since it requires less system knowledge for implementation and reliable data is commonly available in smart engineering systems. Several data-driven predictive control algorithms have been developed recently, which largely follow similar approaches, but with specific formulations and tuning parameters. This review aims to provide a structured and accessible guide on linear data-driven predictive control methods and practices for people in both academia and the industry seeking to approach and explore this field. To do so, we first discuss standard methods, such as subspace predictive control (SPC), and data-enabled predictive control (DeePC), but we also include newer hybrid approaches to DPC, such as $\gamma$–data-driven predictive control and generalized data-driven predictive control. For all presented data-driven predictive controllers we provide a detailed analysis regarding the underlying theory, implementation details and design guidelines, including an overview of methods to guarantee closed-loop stability and promising extensions towards handling nonlinear systems. The performance of the reviewed DPC approaches is compared via simulations on two benchmark examples from the literature, allowing us to provide a comprehensive overview of the different techniques in the presence of noisy data.

This paper considers the problem of distributed motion- and task-planning of multi-agent and multi-agent-object systems under temporal-logic-based tasks and uncertain dynamics. We focus on manipulator-endowed robotic agents that can interact with their surroundings. We present first continuous control algorithms for multi-agent navigation and cooperative object manipulation that exhibit the following properties. First, they are distributed in the sense that each agent calculates its own control signal from local interaction with the other agents and the environment. Second, they guarantee safety properties in terms of inter-agent collision avoidance and obstacle avoidance. Third, they adapt on-the-fly to dynamic uncertainties and are robust to exogenous disturbances. The aforementioned algorithms allow the abstraction of the underlying system to a finite-state representation. Inspired by formal-verification techniques, we use such a representation to derive plans for the agents that satisfy the given temporal-logic tasks. Various simulation results and hardware experiments verify the efficiency of the proposed algorithms.