Annual Reviews in Control最新文献

This paper presents a tutorial overview of path integral (PI) approaches for stochastic optimal control and trajectory optimization. We concisely summarize the theoretical development of path integral control to compute a solution for stochastic optimal control and provide algorithmic descriptions of the cross-entropy (CE) method, an open-loop controller using the receding horizon scheme known as the model predictive path integral (MPPI), and a parameterized state feedback controller based on the path integral control theory. We discuss policy search methods based on path integral control, efficient and stable sampling strategies, extensions to multi-agent decision-making, and MPPI for the trajectory optimization on manifolds. For tutorial demonstrations, some PI-based controllers are implemented in Python, MATLAB and ROS2/Gazebo simulations for trajectory optimization. The simulation frameworks and source codes are publicly available at the github page.

This vision article shows how to build on the framework of event-triggered Control Barrier Functions (CBFs) to design model-free controllers for safety-critical multi-agent systems with unknown dynamics, including humans in the loop. This event-triggered framework has been shown to be computationally efficient and robust while guaranteeing safety for systems with unknown dynamics. We show how to extend it to model-free safety critical control where a controllable ego agent does not need to model the dynamics of other agents and updates its control based only on events dependent on the error states of agents obtained by real-time sensor measurements. To facilitate the process of real-time sensor measurements critical in this approach, we also present CBF relative degree reduction methods, which can reduce the number of such measurements. We illustrate the effectiveness of the proposed framework on a multi-agent traffic merging decentralized control problem and on highway lane changing control with humans in the loop and relative degree reduction. We also compare the proposed event-driven method to the classical time-driven approach.

Inspections of industrial and civil infrastructures prevent unexpected failures that may lead to loss of life. Although inspection robotics is gaining momentum, most of field operations are still performed by human workers. For inspection robots, the main limiting factors are the low versatility and reliability in dynamic, non-structured and highly complex environments. To tackle these issues, we have designed a modular and self-reconfigurable hybrid platform, which consists of three units: the mobile Main Base and two Crawler Units with docking interfaces. The Crawler Unit operates in constrained environments and narrow spaces, while the Main Base will inspect wide areas and deploy/recover the Crawler Units near/from inspection sites, as in marsupial robots. Docking interfaces will allow the Crawler Units to reconfigure into a snake robot or mobile manipulators. In particular, the Crawler Units consist of four modules connected by three kinematic chains for nine active joints in total. Each module is equipped with half active, half passive tracks for moving. This paper discusses in detail the dynamic model of the Crawler Unit, especially focusing on the definition of effective constraint equations, which closely model the system features avoiding common simplifications. Numerical simulations and physical experiments validate the proposed dynamic model of the Crawler Unit.

2D systems, also known as doubly-indexed systems, have gained an increasingly special attention in the control community, as they allow for modeling systems with more complex dynamics than the classical so called 1D systems where the signals are indexed by one variable only usually representing the time. Like for 1D systems, stability conditions have been proposed for 2D systems in the form of a linear matrix inequality (LMI) feasibility test, as such conditions may be tested by solving a convex optimization problem, and as such conditions may open the door for a number of developments such as establishing robust stability and designing stabilizing controllers. This paper aims at presenting, under a unified framework, various LMI stability conditions for 2D systems that have been proposed in the literature, from pioneering to recent contributions, in order to provide the reader with a comprehensive collection that may serve as a source of historical information as well as a platform for comparing the major characteristics of each condition. Also, this paper proposes novel investigations of the presented conditions, in particular through conservatism and complexity analyses carried out in the best cases, in the worst cases, and for various specific numerical examples with different type of dynamics, dimensions and difficulty.

Guidance and control (G&C) technologies play a central role in the development and operation of vehicular systems. The emergence of computational guidance and control (CG&C) and highly efficient numerical algorithms has opened up the great potential for solving complex constrained G&C problems onboard, enabling higher level of autonomous vehicle operations. In particular, convex-optimization-based G&C has matured significantly over the years and many advances continue to be made, allowing the generation of optimal G&C solutions in real-time for many vehicular systems in aerospace, automotive, and other domains. In this paper, we review recent major advances in convex optimization and convexification techniques for G&C of vehicular systems, focusing primarily on three important application fields: (1) Space vehicles for powered descent guidance, small body landing, rendezvous and proximity operations, orbital transfer, spacecraft reorientation, space robotics and manipulation, spacecraft formation flying, and station keeping; (2) Air vehicles including hypersonic/entry vehicles, missiles and projectiles, launch/ascent vehicles, and low-speed air vehicles; and (3) Motion control and powertrain control of ground vehicles. Throughout the paper, we draw figures that illustrate the basic mission concepts and objectives, introduce key equations that characterize the feature of each class of problems and approaches, and present tables that summarize similarities and distinctions among the problems, ideas, and methods. Where available, we provide comparative analyses and reveal correlations between different applications and technical approaches. Finally, we identify open challenges and issues, discuss potential opportunities, and make suggestions for future research directions.

Chaos control remains a crucial area of study in nonlinear dynamics due to its ability to enhance system stability and efficiency in various applications. This review thoroughly examines modern chaos control techniques and offers new insights and methods for stabilizing inherently unpredictable systems. It discusses recent advancements in chaos control, focusing on theoretical breakthroughs and practical applications. Various methods for controlling chaos are explored, including the OGY method, Delayed Feedback Control (DFC), Proportional–Integral–Derivative (PID) control, Sliding Mode Control (SMC), and some unconventional techniques, evaluating their effectiveness in different chaotic systems. By analyzing the literature, this review highlights the potential of chaos control techniques to enhance system predictability and reliability, opening up promising paths for future research.

The use of robots has exceeded the standard focus of manufacturing and production. Over the last decades, special robotic systems have been developed in various extreme environments, such as in the maintenance, repair or even decommissioning of large-scale, strategic facilities, important to any nation’s infrastructure, including power, space, mining, etc. The deployment areas for these robots, like nuclear fuel handling systems, are generally hazardous or unreachable for human beings. The control techniques therein will play an indispensable role in the overall performance of a robotic system as they need to answer enhanced requirements for performance, robustness, and long-term reliability, driven by the fundamental demand for safe operation in complex and hazardous environments. This also needs an understanding of the enhanced industrial standards and requirements for the research, development, design and use of control systems in such environments. The control systems need to be designed specifically capable of tackling different practical control challenges caused by extreme environmental factors. This special section is designed and motivated to bridge the gap between the research community and application engineers, and to help connect control theory, control applications and industrial requirements/regulations.

Operating electrolyzers for producing green hydrogen is a critical emerging issue because of either the broader use of hydrogen for several scopes or the short life span and efficiency of these components. Digital Twin offers a new opportunity to effectively face these problems by improving online control and providing fault detection, diagnosis, and prediction services. Since the Digital Twin is, in fact, a virtual mirror of a real system continuously updated by information received from the field, it allows it to swiftly react to small signals of departure from standard or optimal conditions. Although Digital Twins are widely applied in different fields, comprehensive guidance on developing and designing a Digital Twin in the literature is still lacking. This manuscript aims to provide a comprehensive guide on how to build the Digital Twin of a PEM-Electrolyzer. In particular, the architecture of the Digital Twin is initially presented, then all its components are analyzed, showing the steps to be performed to build a Digital Twin for operating PEM-Electrolyser system.

In this tutorial we study a safety analog of the classical zero-sum differential game with positive definite penalties on the state and the two inputs. Consider a nonlinear system affine in two inputs, which are called “offender” and “defender.” Let the inputs have the opposing objectives in relation to an infinite-time cost which, in addition to penalizing the inputs of both agents, incorporates a safety index of the system (a barrier function), with the defender aiming to maximize the system safety and the offender aiming to minimize it. If there is a pair of (offender, defender) non-Nash feedback policies of the $L_{g}h$ form with a safe outcome, namely, where the defender maintains safety while the offender fails to violate safety, then there exists an inverse optimal pair of policies that attain a Nash equilibrium relative to the safety minimax objective. In the tutorial we study both deterministic and stochastic offenders. The deterministic offender applies its feedback through its deterministic input value, while the stochastic offender applies its feedback through its incremental covariance. In addition to Nash policies for a minimax offender–defender formulation, we provide feedback laws for the defender, in the scenario where the offender action is unrestricted by optimality, and where the defender ensures input-to-state safety in the deterministic and stochastic senses. This tutorial is derived from our recent article on inverse optimal safety filters, by setting the nominal control to zero and declaring the disturbance to be the offender agent.

Among several illustrative examples, one is particularly interesting and unconventional. We consider a safety game played on a unicycle vehicle between its two inputs: the angular velocity and the linear velocity, as the opposing players. We consider two scenarios. In the first, the angular velocity, acting as an offender, attempts to run the vehicle into an obstacle by steering, while the linear velocity, acting as a defender, drives the vehicle forward or in reverse to prevent the vehicle being run into the obstacle. In the second scenario, the linear velocity acts as an offender and angular velocity acts as a defender (in the deterministic case by varying the heading rate; in the stochastic case by varying the variance of a white noise driving the heading rate). A “wind” towards the obstacle advantages the offender in both scenarios. The input policies derived are optimal in the sense of their opposite objectives, under the best possible policy of the opponent, under meaningful costs on their actions. The linear velocity input prevails, whether acting in the role of a defender, in which case the collision with the obstacle is prevented, or in the role of an offender, in which case the collision with the obstacle is achieved.