Current Robotics Reports

Recent advances in sensing, actuation, and computation have opened the door to multi-robot systems consisting of hundreds/thousands of robots, with promising applications to automated manufacturing, disaster relief, harvesting, last-mile delivery, port/airport operations, or search and rescue. The community has leveraged model-free multi-agent reinforcement learning (MARL) to devise efficient, scalable controllers for multi-robot systems (MRS). This review aims to provide an analysis of the state-of-the-art in distributed MARL for multi-robot cooperation. Decentralized MRS face fundamental challenges, such as non-stationarity and partial observability. Building upon the “centralized training, decentralized execution” paradigm, recent MARL approaches include independent learning, centralized critic, value decomposition, and communication learning approaches. Cooperative behaviors are demonstrated through AI benchmarks and fundamental real-world robotic capabilities such as multi-robot motion/path planning. This survey reports the challenges surrounding decentralized model-free MARL for multi-robot cooperation and existing classes of approaches. We present benchmarks and robotic applications along with a discussion on current open avenues for research.

Current real-world interaction between humans and robots is extremely limited. We present challenges that, if addressed, will enable humans and robots to collaborate fluently. Humans and robots have unique advantages best leveraged in Human-Robot Teams. However, human and robot collaboration is challenging, and creating algorithmic advances to support teaming requires careful consideration. Prior research on Human-Robot Interaction, Multi-Agent Robotics, and Human-Centered Artificial Intelligence is often limited in scope or application due to unique challenges in combining humans and robots into teams. Identifying the key challenges that apply to a broad range of Human-Robot Teaming applications allows for a focused and collaborative development of a future toward a world where humans and robots can work together in every layer of society. In order to realize the potential of Human-Robot Teaming while avoiding potential societal harm, several key challenges must be addressed: (1) Communication, (2) Modeling Human Behavior, (3) Long-Term Interaction, (4) Scalability, (5) Safety, (6) Privacy, (7) Ethics, (8) Metrics and Benchmarking, (9) Human Social and Psychological Wellbeing.

Starting with a technical categorization and an overview of current exoskeletons and orthoses and their applications, this review focuses on robotic exoskeletons and orthoses for neuromotor rehabilitation and relevant research needed to provide individualized adaptive support to people under complex environmental conditions, such as assisted daily living. Many different approaches from the field of autonomous robots have recently been applied to the control of exoskeletons. In addition, approaches from the field of brain-computer interfaces for intention recognition are being intensively researched to improve interaction. Finally, besides stimulation, bidirectional feedback and feedback-based learning are recognized as very important to enable individualized, flexible, and adaptive human assistance. AI-based methods for adaptation and online learning of robotic exoskeleton control, combined with intrinsic recognition of human intentions and consent, will in particular lead to improving the quality of human–robot interaction and thus user satisfaction with exoskeleton-based rehabilitation interventions.

The discovery of living organisms under extreme environmental conditions of pressure, temperature, and chemical composition on Earth has opened up the possibility of existence and persistence of life in extreme environment pockets across the solar system. These environments range from the many intriguing moons, to the deep atmospheres of Venus and even the giant gas planets, to the small icy worlds of comets and Kuiper Belt Objects (KBOs). Exploring these environments can ascertain the range of conditions that can support life and can also identify planetary processes that are responsible for generating and sustaining habitable worlds. These environments are also time capsules into early formation of the solar system and will provide vital clues of how our early solar system gave way to the current planets and moons. Over the last few decades, numerous missions started with flyby spacecraft, followed by orbiting satellites and missions with orbiter/lander capabilities. Since then, there have been numerous missions that have utilized rovers of ever-increasing size and complexity, equipped with state-of-the-art laboratories on wheels. Although current generations of rovers achieve mobility through wheels, there are fundamental limitations that prevent these rovers from accessing rugged environments, cliffs, canyons, and caves. These rugged environments are often the first places geologist look to observe stratification from geohistorical processes. There is an important need for new robot mobility solutions, like hopping, rolling, crawling, and walking that can access these rugged environments like cliffs, canyons, and caves. These new generations of rovers have some extraordinary capabilities including being able to grip onto rocks like NASA/JPL LEMUR 2, operate in swarms such as MIT’s microbots, or have high-specific energy fuel cell power supply that is approximately 40-fold higher than conventional lithium ion batteries to Stanford/NASA JPL’s Hedgehog which is able to hop and somersault in low-gravity environments such asteroids. All of these mobility options and supporting technologies have been proposed and developed to explore these hard-to-reach unconventional environments. This article provides a review of the robotic systems developed over the past few decades, in addition to new state-of-the-art concepts that are leading contenders for future missions to explore extreme environments on Earth and off-world.

Robotic patrolling aims at protecting a physical environment by deploying a team of one or more autonomous mobile robots in it. A key problem in this scenario is characterizing and computing effective patrolling strategies that could guarantee some level of protection against different types of threats. This paper provides a survey of contributions that represent the recent research trends to deal with such a challenge. Starting from a set of basic and recurring modeling landmarks, the formulations of robotic patrolling studied by current research are diverse and, to some extent, complementary. Some works propose optimal approaches where the objective function is based on the idleness induced by the patrolling strategy on locations of the environment. On-line methods focus on handling events that can dynamically alter the patrolling task. Adversarial methods, where an underlying game-theoretical interaction with an attacker is modeled, consider sophisticated attacker behaviors. The wide spectrum of heterogenous approaches and techniques shows a common trend of moving towards more realistic models where constraints, dynamic environments, limited attacker capabilities, and richer strategy representations are introduced. The results provide complementarities and synergies towards more effective robotic patrolling systems, paving the way to a set of interesting open problems.

Research in underwater manipulation has mostly focused on solving individual parts of the manipulation challenge; however, we believe a systemic approach needs to be taken to achieve full autonomy. With this survey, we aim to provide a review of the different dynamic modeling, control, motion planning, and perception methodologies presented in the literature, and, more importantly, we intend to highlight the necessary steps that need to be taken to achieve fully autonomous underwater manipulation. Achieving autonomous manipulation in underwater environments is a complex and multi-disciplinary challenge. Recent works have focused on moving from simulation-based environments to experimental validation of the proposed methods. Furthermore, the advancements of machine learning have been making an impact in the underwater manipulation, data-driven strategies playing a central role in the last years developments. We present an overview of the current trends in the area of autonomous underwater manipulation. First, we provide a review of state-of-the-art algorithms developed in the area of dynamic modeling, control, motion planning, and perception. Second, we discuss the limitations of the current systems and present possible avenues to obtain robust autonomous manipulation.

This review provides an overview of the state of the art in bioinspired soft robotics by examining advancements in actuation, functionality, modeling, and control. Recent research into actuation methods, such as artificial muscles, has expanded the functionality and potential use of bioinspired soft robots. Additionally, the application of finite dimensional models has improved computational efficiency for modeling soft continuum systems, and garnered interest as a basis for controller formulation. Bioinspiration in the field of soft robotics has led to diverse approaches to problems in a range of task spaces. In particular, new capabilities in system simplification, miniaturization, and untethering have each contributed to the field’s growth. There is still significant room for improvement in the streamlining of design and manufacturing for these systems, as well as in their control.

Humanoid robots are versatile platforms with the potential to assist humans in several domains, from education to healthcare, from entertainment to the factory of the future. To find their place into our daily life, where complex interactions and collaborations with humans are expected, their social and physical interaction skills need to be further improved. The hallmark of humanoids is their anthropomorphic shape, which facilitates the interaction but at the same time increases the expectations of the human in terms of advanced cooperation capabilities. Cooperation with humans requires an appropriate modeling and real-time estimation of the human state and intention. This information is required both at a high level by the cooperative decision-making policy and at a low level by the interaction controller that implements the physical interaction. Real-time constraints induce simplified models that limit the decision capabilities of the robot during cooperation. In this article, we review the current achievements in the context of human-humanoid interaction and cooperation. We report on the cognitive and cooperation skills that the robot needs to help humans achieve their goals, and how these high-level skills translate into the robot’s low-level control commands. Finally, we report on the applications of humanoid robots as humans’ companions, co-workers, or avatars.