Prefacedoi: 10.1088/1757-899x/1261/1/011001pmid: N/A
Whilst providing a universal, clear and meaningful definition of ‘Embodied Intelligence’ is challenging, there is however a clear direction and motivation of the research community to further our understanding of systems that exhibit embodied intelligence and to understand how it can be designed and leveraged. This is a challenging topic, and requires an interdisciplinary and cooperative approach from researchers from engineering, social sciences, computer sciences, biological, materials science and more.Furthering our understanding or Embodied Intelligence requires advancement in the philosophical and theoretical domains to provide frameworks intelligence, consciousness and learning. However the inherently physicality of embodied intelligence relies on the creation of physical technologies and real world interactions, requiring increasingly smart and novel technologies. The intersection of these domains, in a world where AI and machine intelligence are playing an ever growing role in our daily lives makes this research field timely, fascinating and significant.In this book we provide a representative collection of papers written by the leading researchers in the field who attended and participated in the Embodied Intelligence Workshop, held virtually in March 2021. This includes short reviews, perspectives and original research articles. These aim to provide thought provoking, interdisciplinary and novel views and perspectives on this field, its future direction and some of the key concepts needed to deliver breakthroughs in our understanding or practical implementation.The book first touches on perspectives and outlooks for Embodied Intelligence before including examples of soft and bio-inspired robotics. This is followed by chapters discussing the principles of Embodied Intelligence in robotics and also examples of embodiment for real-world applications. Finally, the last two chapters focus on cognition and learning and also the philosophical and conceptual issues surrounding Embodied Intelligence.We would like to thank all of the participants of the seminar, the authors, and the reviewers for their excellent contributions to this volume.
Benefits and Challenges of Morphological Adaptation in Real-world RobotsNygaard, Tønnes F.; Glette, Kyrre
doi: 10.1088/1757-899x/1261/1/012017pmid: N/A
For many years, the main target of robotics research has been how to improve control and behavior to allow robots to solve more challenging tasks. Recently, the concept of embodied intelligence has shifted this focus from pure control to a more holistic approach. The mind, body, environment, and the interactions between all these are all highly important for the performance of a robot system. This has previously been investigated in virtual robots in simulation or very simple physical robots in the lab, but the benefits for more advanced and capable real-world robots are still to be determined. We present a case study with several examples of real-world adaptation of morphology, and discuss the benefits and challenges as this direction becomes more important in the field of robotics. Taking an embodied approach promises more adaptive and robust robots, but there are many challenges that need to be addressed before wide-spread adoption in the robotics community.
Embodiment in Dialogue: Daily Dialogue Android Based on Multimodal InformationUchida, Takahisa; Minato, Takashi; Ishiguro, Hiroshi
doi: 10.1088/1757-899x/1261/1/012016pmid: N/A
The research and development of robots that can have daily dialogues autonomously with humans have become increasingly important. Most of them, however, have been limited to studies on speech recognition and interfaces. As the result, they still cannot build relationships with humans through long-term daily dialogue. In this background, we have developed a robot that can understand human relationships with their intentions and desires, and can interact with them in human-like ways through various modalities. In this chapter, we first describe an android platform for multimodal interaction based on intentions and desires. Here, both the user’s and the robot’s subjective preferences and opinions are essential. Then, we introduce a function to model the user’s preference, which is necessary for building human relationships. We also discuss the relationship between dialogue and the opinions that can be attributed to the android considering its embodiment. Finally, based on the above studies, we discuss the embodiment in dialogue and future work.
Morphological Computation and Control ComplexityThuruthel, Thomas George; Iida, Fumiya
doi: 10.1088/1757-899x/1261/1/012011pmid: N/A
Morphological computation proposes the idea that in a physical system, certain computational processes can be off-loaded to the body. However, the concept has still eluded serious theoretical quantification attempts, unlike traditional computational theory. This perspective examines the notion of morphological computation from the well established theories of traditional computation and computational complexity, drawing parallels between the two, to understand the differences and similarities. Further, we look at the quantification efforts of morphological computation and attempt to link it to the unexplored field of control complexity. We argue that the development of complexity theory for control problems is necessary to study and utilize the concept of morphological computation, if it is possible.
Peer Review Statementdoi: 10.1088/1757-899x/1261/1/011002pmid: N/A
All papers published in this volume have been reviewed through processes administered by the Editors. Reviews were conducted by expert referees to the professional and scientific standards expected of a proceedings journal published by IOP Publishing Publishing.• Type of peer review: Single Anonymous• Conference submission management system: Morressier• Number of submissions received: 27• Number of submissions sent for review: 24• Number of submissions accepted: 27• Acceptance Rate (Submissions Accepted / Submissions Received × 100): 100• Average number of reviews per paper: 2.2• Total number of reviewers involved: 33• Contact person for queries:Name: Josie HughesEmail: [email protected]: EPFL - IGM
Multifunctional Underwater Soft Robots: A Simulation EssayMathew, Anup Teejo; Armanini, Costanza; Alshehhi, Aysha Ali Samra Ali; Hmida, Ikhlas Mohamed Ben; Renda, Federico
doi: 10.1088/1757-899x/1261/1/012008pmid: N/A
Underwater soft robotics is receiving growing popularity within the scientific community, thanks to its prospective capability of tackling challenges that are hard to deal with using traditional rigid technologies, especially while interacting with an unstructured environment. Recently, we proposed a multi-module underwater robotic system with deformable propellers, inspired by bacteria morphology [1]. Here, the same bio-inspired modular structure is employed to perform manipulation tasks, in order to design a multi-functional integrated system. Employing the Geometric Variable Strain Approach, we simulate a scenario where the flagellated robot moves towards a preferred target and, using the same soft appendages, it hooks to it, simulating a monitoring task. The modeling approach and the design allow the Embodied Intelligence principles to exploit the robot’s surrounding environment (water), the shape of the grip-target and the robot’s compliant nature to mediate effective navigation and safe interaction with the target, using few control inputs.
Mechanical Sensing in Embodied AgentsPreti, Matteo Lo; Thuruthel, Thomas George; Gilday, Kieran; Beccai, Lucia; Iida, Fumiya
doi: 10.1088/1757-899x/1261/1/012013pmid: N/A
Sensors enable autonomous systems to obtain information about their internal states and the environment for guiding their actions. It is as essential for these sensors to reject disturbances as to gather the correct information. There are numerous trade-offs and considerations in designing these sensory systems. For instance, natural agents evolved a vast diversity of highly optimized sensory organs to perform their tasks. This work focuses on how these sensory systems estimate mechanical stimuli. We look at some of the strategies and design principles found in nature to understand fundamental trade-offs and design considerations when acquiring and processing mechanical information.
Physical Intelligence in BiomechanicsWang, Shiqiang; Shi, Yongqi; Wen, Li
doi: 10.1088/1757-899x/1261/1/012012pmid: N/A
The intelligence of the biological agents is enabled by their neural intelligence in their brains, in the meantime, their physical intelligence encoded in the bodies plays a nonnegligible role. The development of intelligent bio-inspired robots depends on an in-depth understanding of physical intelligence, particularly in biomechanics. Here, we briefly review the physical intelligence of biological organisms from three aspects: material, structure, and morphology. We envision that bio-inspired physical intelligence would boost the development of future intelligent lifelike machines.
Active inference, preference learning and adaptive behaviourSajid, Noor; Tigas, Panagiotis; Friston, Karl
doi: 10.1088/1757-899x/1261/1/012020pmid: N/A
The ability to adapt to a changing environment underwrites sentient behaviour e.g., wearing a raincoat when walking in the rain but removing it when indoors. In such instances, agents act to satisfy some preferred mode of behaviour that leads to predictable states necessary for survival, i.e., states that are characteristic of that agent. In this chapter, we describe how active inference agents, equipped with preference learning, can exhibit these distinct behavioural modes – influenced by environment dynamics – to aptly trade-off between preference satisfaction and exploration. We validate this in a modified OpenAI Gym FrozenLake environment (without any extrinsic signal) with and without volatility under a fixed model of the environment. In a static (i.e., without volatility) environment, preference-learning agents accumulate confident (Bayesian) beliefs about their behaviour and act to satisfy them. In contrast, volatile dynamics led to preference uncertainty and exploratory behaviour. This demonstrates that active inference agents, equipped with preference learning, have the appropriate machinery to (i) engage in adaptive behaviour under appropriate levels of volatility, and (ii) learn context-dependent subjective preferences.
Coevolution of internal representations in physical human-robot orchestration – models of the surgeon and the robot in robotic surgeryNisky, Ilana; Costi, Leone; Iida, Fumiya
doi: 10.1088/1757-899x/1261/1/012014pmid: N/A
In teleoperated Robot-Assisted Minimally-Invasive Surgery (RAMIS), a surgeon controls the movements of instruments inside the patient’s body via a pair of robotic joysticks. RAMIS has transformed many surgical disciplines, but its full potential is still to be realized. In this chapter we propose a pathway towards overcoming several bottlenecks that are related to transparency and stability of the teleoperation channels that mediate RAMIS. We describe the traditional system centered and the more recent human-centred approaches to teleoperation, and the special considerations for RAMIS as an application of teleoperation. However, the human-centered approach is still one sided view focusing on the surgeon but neglecting the learning capabilities of robotic systems. Hence, we consider a more general idea of physical human-robot orchestration with coevolution of mutual internal representations – of the human and the robot, and discuss it in comparison to human-human collaboration over teleoperated channels.