Annual Review of Control, Robotics, and Autonomous Systems - Current Issue

Focus on the ethics of a given technology tends to lag far behind its development. This lag has been particularly acute in the case of artificial intelligence, whose accelerated deployment in a wide range of domains has triggered unprecedented attention on the risks and consequences for society at large, leading to a myriad of ethics regulations, which are difficult to coordinate and integrate due to their late appearance. The very nature of social robots forces their deployment to occur at a much slower pace, providing an opportunity for a profound reflection on ethics, which is already happening in multidisciplinary teams. This article provides a personal view of the ethics landscape, centered on the particularities of social robotics, with the main issues being ordered along two axes (individual and societal) and grouped into eight categories (human dignity, human autonomy, robot transparency, emotional bonding, privacy and safety, justice, freedom, and responsibility). This structure stems from the experience of developing and teaching a university course on ethics in social robotics, whose pedagogical materials are freely available.

A multiagent system should be capable of fast and flexible decision-making to successfully manage the uncertainty, variability, and dynamic change encountered when operating in the real world. Decision-making is fast if it breaks indecision as quickly as indecision becomes costly. This requires fast divergence away from indecision in addition to fast convergence to a decision. Decision-making is flexible if it adapts to signals important to successful operation, even if they are weak or rare. This requires tunable sensitivity to input for modulating regimes in which the system is ultrasensitive and in which it is robust. Nonlinearity and feedback in the decision-making process are necessary to meeting these requirements. This article reviews theoretical principles, analytical results, related literature, and applications of decentralized nonlinear opinion dynamics that enable fast and flexible decision-making among multiple options for multiagent systems interconnected by communication and belief system networks. The theory and tools provide a principled and systematic means for designing and analyzing decision-making in systems ranging from robot teams to social networks.

Recent years have seen significant progress in the realm of robot autonomy, accompanied by the expanding reach of robotic technologies. However, the emergence of new deployment domains brings unprecedented challenges in ensuring safe operation of these systems, which remains as crucial as ever. While traditional model-based safe control methods struggle with generalizability and scalability, emerging data-driven approaches tend to lack well-understood guarantees, which can result in unpredictable catastrophic failures. Successful deployment of the next generation of autonomous robots will require integrating the strengths of both paradigms. This article provides a review of safety filter approaches, highlighting important connections between existing techniques and proposing a unified technical framework to understand, compare, and combine them. The new unified view exposes a shared modular structure across a range of seemingly disparate safety filter classes and naturally suggests directions for future progress toward more scalable synthesis, robust monitoring, and efficient intervention.

Researchers in human–robot collaboration have extensively studied methods for inferring human intentions and predicting their actions, as this is an important precursor for robots to provide useful assistance. We review contemporary methods for intention inference and human activity prediction. Our survey finds that intentions and goals are often inferred via Bayesian posterior estimation and Markov decision processes that model internal human states as unobserved variables or represent both agents in a shared probabilistic framework. An alternative approach is to use neural networks and other supervised learning approaches to directly map observable outcomes to intentions and to make predictions about future human activity based on past observations. That said, due to the complexity of human intentions, existing work usually reasons about limited domains, makes unrealistic simplifications about intentions, and is mostly constrained to short-term predictions. This state of the art provides opportunity for future research that could include more nuanced models of intents, reason over longer horizons, and account for the human tendency to adapt.

Operating autonomous agents in unstructured space presents a difficult problem. The complexity of making decisions such as when to yield and when to go ahead increases exponentially with the number of agents. This is true for humans as well as for software that controls autonomous agents. With some practice, however, human operators are able to move efficiently in a maze of interacting agents in dense traffic. One recent result correlates the instability of equilibria in a multiagent system with an absence of gridlocks. These control barrier function–based algorithms do not include a decision-making component—the action is continuous, and negotiation happens through instability. This mechanism, referred to as instinctive negotiation, is contrasted with discontinuity-induced decisions arising from nonconvex optimization. Based on observed behavioral similarities and insights into human implicit and explicit learning, this article proposes a connection with human driving and suggests that humans may employ a mechanism similar to instinctive negotiation to navigate dense traffic.

More than 1 billion people in the world are estimated to experience significant disability. These disabilities can impact people's ability to independently conduct activities of daily living, including ambulating, eating, dressing, taking care of personal hygiene, and more. Mobile and manipulator robots, which can move about human environments and physically interact with objects and people, have the potential to assist people with disabilities in activities of daily living. Although the vision of physically assistive robots has motivated research across subfields of robotics for decades, such robots have only recently become feasible in terms of capabilities, safety, and price. More and more research involves end-to-end robotic systems that interact with people with disabilities in real-world settings. In this article, we survey papers about physically assistive robots intended for people with disabilities from top conferences and journals in robotics, human–computer interactions, and accessible technology, to identify the general trends and research methodologies. We then dive into three specific research themes—interaction interfaces, levels of autonomy, and adaptation—and present frameworks for how these themes manifest across physically assistive robot research. We conclude with directions for future research.

The integration of more artificial intelligence (AI)–enabled robots for the military and first response domains is necessary to support long-duration deployments in uncertain and dynamic environments while lessening humans’ exposure to threats and dangers. The effective integration of AI-enabled robots as teammates with humans will provide support and enhance overall mission performance; however, the majority of current research on human–robot interaction focuses only on the robot team supervisor. The true integration of robots into military and first response missions will require a breadth of human roles that span from the highest command level to the dismounted in situ personnel working directly with robots. All human roles within the hierarchy must understand and maintain direct control of the robot teammates. This article maps existing human roles from the literature to a military mission, presents technical challenges associated with this future human–robot teaming, and provides potential solutions and recommendations to propel the field forward toward human–robot teams that can achieve domain-relevant missions.

This review deals with the control of parabolic trough collector (PTC) solar power plants. After a brief introduction, we present a description of PTC plants. We then provide a short literature review and describe some of our experiences. We also describe new control trends in PTC plants. Recent research has focused on (a) new control methods using mobile sensors mounted on drones and unmanned ground vehicles as an integral part of the control systems; (b) spatially distributed solar irradiance estimation methods using a variable fleet of sensors mounted on drones and unmanned ground vehicles; (c) strategies to achieve thermal balance in large-scale fields; (d) new model predictive control algorithms using mobile solar sensor estimates and predictions for safer and more efficient plant operation, which allow the effective integration of solar energy and combine coalitional and artificial intelligence techniques; and (e) fault detection and diagnosis methods to ensure safe operation.

Wind energy is recognized worldwide as cost-effective and environmentally friendly, and it is among the fastest-growing sources of electrical energy. To further decrease the cost of wind energy, wind turbines are being designed at ever-larger scales. To expand the deployment of wind energy, wind turbines are also being designed on floating platforms for placement in deep-water locations offshore. Both larger-scale and floating wind turbines pose challenges because of their greater structural loads and deflections. Complex, large-scale systems such as modern wind turbines increasingly require a control co-design approach, whereby the system design and control design are performed in a more integrated fashion. This article reviews recent developments in control co-design of wind turbines. We provide an overview of wind turbine design objectives and constraints, issues in the design of key wind turbine components, modeling of the wind turbine and environment, and controller coupling issues. Wind turbine control functions and the integration of control design in co-design are detailed with a focus on co-design compatible control approaches.

The control of wave energy converters (WECs) to maximize power capture is a challenging problem. In particular, the nature of the wave excitation, which is in general panchromatic (or multi-sinusoidal), presents a reciprocating energy source that needs to be rectified through some means. In addition, the development of suitable control-oriented models is also challenging, requiring correct representation of system hydrodynamics and power take-off (PTO) components, while also lending themselves to control synthesis and real-time computational performance, along with a challenging optimal control problem. This article presents a moment-based mathematical framework for the formulation and solution of WEC control. It shows that moments are ideally suited to WEC control in terms of their ability to accurately characterize the nature of the wave excitation force (and the consequent evolutions in the system variables) while also gracefully including hydrodynamic and PTO nonlinearities as well as a natural extension to WEC arrays. Model reduction, to mold the system model into a control-friendly form, is also a feature of this framework.

Robotics is powered by software. Software tools control the rate of innovation in robotics research, drive the growth of the robotics industry, and power the education of future innovators and developers. Nearly 900,000 open-source repositories on GitHub are tagged with the keyword robotics—a potentially vast resource, but only a fraction of those are truly accessible in terms of quality, licensability, understandability, and total cost of ownership. The challenge is to match this resource to the needs of students, researchers, and companies to power cutting-edge research and real-world industrial solutions. This article reviews software tools for robotics, including both those created by the community at large and those created by the authors, as well as their impact on education, research, and industry.

Sampling-based motion planning is one of the fundamental paradigms to generate robot motions, and a cornerstone of robotics research. This comparative review provides an up-to-date guide and reference manual for the use of sampling-based motion planning algorithms. It includes a history of motion planning, an overview of the most successful planners, and a discussion of their properties. It also shows how planners can handle special cases and how extensions of motion planning can be accommodated. To put sampling-based motion planning into a larger context, a discussion of alternative motion generation frameworks highlights their respective differences from sampling-based motion planning. Finally, a set of sampling-based motion planners are compared on 24 challenging planning problems in order to provide insights into which planners perform well in which situations and where future research would be required. This comparative review thereby provides not only a useful reference manual for researchers in the field but also a guide for practitioners to make informed algorithmic decisions.

In recent years, the increasing complexity and safety-critical nature of robotic tasks have highlighted the importance of accurate and reliable robot models. This trend has led to a growing belief that, given enough data, traditional physics-based robot models can be replaced by appropriately trained deep networks or their variants. Simultaneously, there has been a renewed interest in physics-based simulation, fueled by the widespread use of simulators to train reinforcement learning algorithms in the sim-to-real paradigm. The primary objective of this review is to present a unified perspective on the process of determining robot models from data, commonly known as system identification or model learning in different subfields. The review aims to illuminate the key challenges encountered and highlight recent advancements in system identification for robotics. Specifically, we focus on recent breakthroughs that leverage the geometry of the identification problem and incorporate physics-based knowledge beyond mere first-principles model parameterizations. Through these efforts, we strive to provide a contemporary outlook on this problem, bridging classical findings with the latest progress in the field.

This article reviews recent advances in intelligent robotic ultrasound imaging systems. We begin by presenting the commonly employed robotic mechanisms and control techniques in robotic ultrasound imaging, along with their clinical applications. Subsequently, we focus on the deployment of machine learning techniques in the development of robotic sonographers, emphasizing crucial developments aimed at enhancing the intelligence of these systems. The methods for achieving autonomous action reasoning are categorized into two sets of approaches: those relying on implicit environmental data interpretation and those using explicit interpretation. Throughout this exploration, we also discuss practical challenges, including those related to the scarcity of medical data, the need for a deeper understanding of the physical aspects involved, and effective data representation approaches. We conclude by highlighting the open problems in the field and analyzing different possible perspectives on how the community could move forward in this research area.

Considering microbotics, microforce sensing, their working environment, and their control architecture together, microrobotic force-sensing systems provide the potential to outperform traditional stand-alone approaches. Microrobotics is a unique way for humans to control interactions between a robot and micrometer-size samples by enabling the control of speeds, dynamics, approach angles, and localization of the contact in a highly versatile manner. Many highly integrated microforce sensors attempt to measure forces occurring during these interactions, which are highly difficult to predict because the forces strongly depend on many environmental and system parameters. This article discusses state-of-the-art microrobotic systems for microforce sensing, considering all of these factors. It starts by presenting the basic principles of microrobotic microforce sensing, robotics, and control. It then discusses the importance of microforce sensor calibration and active microforce-sensing techniques. Finally, it provides an overview of microrobotic microforce-sensing systems and applications, including both tethered and untethered microrobotic approaches.

Endovascular techniques have revolutionized the treatment of cerebrovascular disease in the human brain. In this review, we examine the current state of this technology, which consists of multiple concentric catheters that are manually navigated from the lumen of peripheral arterial access within the patient's arm or leg up into the brain using fluoroscopic image guidance. There is tremendous potential for the fields of robotics, materials science, and computer science to redefine the current techniques and ultimately improve the safety and efficacy of treatments.

This article makes the case that a powerful new discipline, which we term perception engineering, is steadily emerging. It follows from a progression of ideas that involve creating illusions, from historical paintings and film to modern video games and virtual reality. Rather than creating physical artifacts such as bridges, airplanes, or computers, perception engineers create illusory perceptual experiences. The scope is defined over any agent that interacts with the physical world, including both biological organisms (humans and animals) and engineered systems (robots and autonomous systems). The key idea is that an agent, called a producer, alters the environment with the intent to alter the perceptual experience of another agent, called a receiver. Most importantly, the article introduces a precise mathematical formulation of this process, based on the von Neumann–Morgenstern notion of information, to help scope and define the discipline. This formulation is then applied to the cases of engineered and biological agents, with discussion of its implications for existing fields such as virtual reality, robotics, and even social media. Finally, open challenges and opportunities for involvement are identified.