.
The intention of this e-book was to collect opinions and insights on morphological computation from researchers from a broad range of research as well cultural backgrounds. Instead of the formal style of publications of scientific results, we wanted to provide a rather informal environment where people can present their point of view on morphological computation, its future trajectory and its possibly far-reaching implications. The book is meant to be an inspirational resource and it should allow for cross-fertilization between different disciplines. We also wanted to make sure that the contributions are written in a style accessible for a broad audience.
Please feel free to share this e-book!
The editors: Helmut Hauser, Rudolf M. Füchslin, and Rolf Pfeifer
Chapter 1: Extracting the Full Power of Morphological Computation: Lessons from Case Studies of Robots under Decentralized Control
Abstract: The concept of morphological computation could be the key to under- standing animals’ adaptive behavior. Despite its appeal, its mechanism and application method remain elusive. In this article, we consider how morphological computation can be used effectively by taking some of our studies as practical examples. Through these case studies, we show that the concept of morphological computation can be exploited effectively when we integrate it into autonomous decentralized control systems. We also discuss issues requiring further consideration for the effective use of morphological computation.
Author: Akio Ishiguro
Chapter 2: A Reservoir Computing View of Morphological Computation
Abstract: Morphological Computation is a broad concept that tries to connect the body (or mechanics), brain (or controller) and environment. In this work, we analyze Physical Reservoir Computing as a practical approach to Morphological Computation. Physical Reservoir Computing extends the Reservoir Computing framework that was originally developed in the context of artificial neural networks to physical systems. This has approach provides simple, yet efficient methods to exploit the computational power available in physical dynamical systems.
Authors: Ken Caluwaerts and Benjamin Schrauwen
Chapter 3: Deep into Morphology: Emotions and Functional Structure
Abstract: In this chapter we explore the question of how to scale up the phenomenon of morphological computation to complex adaptive systems. We suggest that an important inspiration may come from the concept of embodied appraisal in human emotion: how physiological states may come to "represent" qualitative features of the interaction. Emotion-based robotics and morphological computation therefore share a common ground, which we explore in this chapter, and venture that insights from each area may prove vital for the development of the other. Our aim is to explore and clarify the potential of this cross-fertilization.
Author: Carlos Herrera- Pérez and Ricardo Sanz
Chapter 4: Zen, Robotics and the Art of Pushing Swings
Abstract: In this short text we argue that a thorough understanding of the principles underlying the simple act of pushing a small child on a swing might shed light on a wide range of phenomena from brain planning of motion to regulation processes in the cell. Those principles can be summarized by what we call “morphological computation.
Author: Fabio Bonsignorio
Chapter 5: A Virtual Material Approach to Morphological Computation
Abstract: In the natural world computation is a matter of survival. Organisms must utilise their own resources in the most efficient way to exploit their environment for food, shelter and reproduction opportunities. Adaptation is achieved at both the individual and population level within a changing environment. For the giant single celled organism slime mould Physarum polycephalum this adaptation is literal: changing its entire body plan during growth, movement, foraging, feeding, and hazard avoidance. Slime mould is remarkable because, despite possessing no nervous system, it has been shown to perform remark- able feats of computation. Slime mould may thus be considered as a living material form of morphological computation. We explore how the concept of material computation by morphological adaptation can be extracted from the behaviour of slime mould to develop distributed multi-agent collectives with emergent, quasi-material behaviour. The natural pattern formation and net- work minimisation of this virtual material may be influenced by the application of environmental stimuli to perform useful computation. We give a brief overview of the approach and demonstrate how the idea of virtual material adaptation can be used for spatially represented unconventional computation and distributed robotics tasks. We conclude by examining the possible future roles and challenges facing material based morphological computation.
Author: Jeff D. Jones
Chapter 6: A Review of Morphological Computation from a Perspective of Heterarchy
Abstract: In the existing contributions to morphological computation, it is found that the morphological property should play a role in interfacing the internal control mechanism of a system with the outside environment. This property not only gives the connection between them, but also mediates, reconciles, and integrates them. However, the problem of what distinguishes the morphological property of the system from that of usual complex control systems seems to remain. As a possible solution, we consider the notion of heterarchy, which is the interplay between hierarchies such as agent and environment. Our discussion reveals that the properties of heterarchical and morphological systems are closely related to each other. For the development of morphological computation, some investigations of the heterarchical property are introduced, and some proposals are submitted.
Author: Kazuto Sasai
Chapter 7: Morphological Computation and Heuristic Bio-Robotics
Abstract: Heuristic Bio-Robotics is a new paradigm for understanding and reproducing animals’ adaptive behavior. In this approach, Morphological Computation is a very important design principle that provides quick response, reduction of formal computation, and adequate coordination transformations for multiple tasks and open environment.
Author: Koh Hosoda
Chapter 8: Morphology: A Concrete Form of Intelligence
Abstract: This article contains our perceptions and comments on the literature concerning morphological computation. We argue that, if there is no computation without intelligence and if morphology is the embodiment of this intelligence, the usage of the term ”morphological computation” is trivial. Therefore, another term should be used to define this new discipline, encompassing all its aspects. We have come up with a new term, ”Morphological Intelligence,” and have identified three subfields, namely, ”Morphological compatibility”, ”Morphological intelligence trans- formation,” and ”Morphological intelligence optimization.” We believe that the ramifications of this new terminology will help us better understand the discipline. In this paper, we explain this concept, using our new terminology, in the context of locomotion.
Authors: Murat Reis and Cihat Ensarioglu
Chapter 9: Molecules and Robots
Abstract: This paper discusses morphological computation from the perspective of self-assembly. Ultimately, organic molecules are “computational machines” that mechanically conduct logical operations and achieve massive structure constructions. Their capability to realize global functionalities by locally interacting with other molecules through their body may endow artificial products with a power of vitality and higher intelligence. As biological molecules attain some fundamental vital activities, such as development or self-repair, we believe that the challenge on filling the gap between artificial machines and organic machines will provide us with hints on what it means to be a living system. This paper discusses morphological computation from the perspective of self-assembly. Ultimately, organic molecules are “computational machines” that mechanically conduct logical operations and achieve massive structure constructions. Their capability to realize global functionalities by locally interacting with other molecules through their body may endow artificial products with a power of vitality and higher intelligence. As biological molecules attain some fundamental vital activities, such as development or self-repair, we believe that the challenge on filling the gap between artificial machines and organic machines will provide us with hints on what it means to be a living system.
Author: Shuhei Miyashita
Chapter 10: Morphological Computation: A Perspective Based on Bacterial Movement
Abstract: There has been a growing interest in the idea and conception of morphological computation and how they may bring benefits to robotics research. This manuscript aims to discuss them based on the movement mechanism of the simplest creatures: bacteria. To be more specific, we will discuss the concept of what morphological computation is by describing the way bacteria connect their “brain,” body, and environment while performing a chemotaxis behavior or random walk. In con- junction with the discussion, we will also suggest some challenges and the necessary future works related with the idea of morphological computation.
Authors: Surya G. Nurzaman, Yoshio Matsumoto, Yutaka Nakamura, Kazumichi Shirai, Satoshi Koizumi, Fumiya Iida, and Hiroshi Ishiguro
Chapter 11: Morphological Computation with Hydrostatic Interaction between Mechanosensory Oscillators
Abstract: Morphological computation is a powerful design concept for coordinating forces from actuators of robots and their environment using their body dynamics (e.g., its geometric structure and viscoelastic distribution) to generate adaptive behavior. However, these forces are invisible to robot designers and therefore very difficult to handle, especially with an external force disturbance. To alleviate this, we focused on the plasmodium of true slime mold as a platform to discuss physical interaction between body parts, because the force balance between its body parts is considered protoplasmic streaming. By exploiting the physical interaction (or so- called morphological computation) between the body parts, the plasmodium exhibits versatile and adaptive behaviors. Inspired by this, we design a hydrostatically inter- acting mechanosensory oscillators. The numerical experiments show that the model produces situation-dependent oscillatory patterns with two modules, and versatile oscillatory patterns with three modules without changing any parameters of the model during the simulation run. The results indicate that embedding morphological computation as a part of control system of robots can contribute to their versatile and adaptive behaviors.
Authors: Takuya Umedachi and Akio Ishiguro
Chapter 12: Evolving Morphological Computation
Abstract: It has been argued that a robot’s morphology (rather than its controller) may “compute.” We hypothesize that there may be circumstances under which there is some advantage for a robot to compute using its body rather than its brain. If this is true, and if we use an evolutionary algorithm to improve the bodies and brains of robots under these circumstances, it should sometimes discover morphological computation and make use of it. Here we argue that morphological complexity may be correlated with morphological computation, and demonstrate a system in which morphological complexity evolves. We also hypothesize about how such a tool could be used to investigate how and when morphological computation is useful.
Author: Josh Bongard
Chapter 13: A Morphological Viewpoint: Juxtaposition of Design Approaches for Locomotion–Rehabilitation Robotics
Abstract: The Morphological Viewpoint: a morphological computation or control system is one which is designed from a morphological point of view.
Authors: Kenneth J. Hunt and Juan Fang
Chapter 14: Morphological Computing and Design
Abstract: A design perspective is taken towards understanding the field of morphological computing from a theoretical and practical perspective. A set of design principles is explored through a chemical morphological computing platform, whereby the physical properties of substances such as, oil droplets, directly inform design tac- tics. These are differently framed to the modes of thought that shape our thinking about machine interactions by using the perspective of process philosophy and its technical embodiment – the assemblage operating system. Specifically, the Bütschli system, which self assembles from the addition of strong alkali to an olive oil field, is considered as a programmable soft robot that can be guided using internal and external chemical cues. This model system is used to explore the metaphysical, material, design and technical challenges in designing with morphological computing where a range of potential application is also discussed ranging from the cybernetic Hylozoic Ground installation for the 2010 Venice Architecture Biennale and the architectural project Future Venice that proposes to direct the action of programmable droplets to grow an artificial reef under the city that stand on narrow wood piles. The aim being to spread its point load and stop it from sinking so quickly into the soft delta soils on which the city has been founded by transforming the stiletto heels of the current foundations into platform boots.
Author: Rachel Armstrong
Chapter 15: Morphological Computation at the Molecular Scale
Abstract: This short essay introduces and states my opinion on future perspectives of morphological computation from the perspective of molecular computing. An analysis of morphological computation is introduced mainly from a viewpoint of “scales” and “interactions.” Some examples of biomolecular computing, especially DNA computing, are briefly described in terms of morphological computation. More- over, possible future connection and possibilities between morphological computation and a new field called “molecular robotics” are mentioned.
Author: Hamada Shogo
Chapter 16: Morphological Computation – A Broad Perspective
Abstract: In this contribution a broad perspective on morphological computation is introduced. We shall argue that shape is a special feature of spatial relations and shall advocate a computation model that takes into account spatial relationships between components of the computation.
Author: Wolfgang Banzhaf
Chapter 17: Trade-Offs in Exploiting Body Morphology for Control: from Simple Bodies and Model-Based Control to Complex Bodies with Model-Free Distributed Control Schemes
Abstract: Tailoring the design of robot bodies for control purposes is implicitly performed by engineers, however, a methodology or set of tools is largely absent and optimization of morphology (shape, material properties of robot bodies, etc.) is lag- ging behind the development of controllers. This has become even more prominent with the advent of compliant, deformable or "soft" bodies. These carry substantial potential regarding their exploitation for control – sometimes referred to as "morphological computation" in the sense of offloading computation needed for control to the body. Here, we will argue in favour of a dynamical systems rather than computational perspective on the problem. Then, we will look at the pros and cons of simple vs. complex bodies, critically reviewing the attractive notion of “soft” bodies automatically taking over control tasks. We will address another key dimension of the design space – whether model-based control should be used and to what extent it is feasible to develop faithful models for different morphologies.
Authors: Matej Hoffmann and Vincent C. Müller
Chapter 18: Morphological Control as Guiding Principle in Physiology and Medical Applications
Abstract: Morphological computation can be loosely defined as the exploitation of the shape, material properties, and intrinsic dynamics of a physical system in order to improve the efficiency of a computation. Morphological control is the application of morphological computing to a control task. In this paper, we discuss possible applications of the concept of morphological control to problems in medicine and clinical therapy. We motivate our conviction that the emergent dynamical systems studied in the various branches of the omics-sciences should be analyzed from the perspective of a broader notion of control engineering and from what has been called biochemical IT (by “omics”-sciences we mean genomics, proteomics, metabolomics, glycomics etc. For the interpretation relevant to this paper, see [23]). This conviction is illustrated by three hypotheses. The first one claims the often experienced age-related loss of control over basic movement patterns, such as walking, are not (only) caused by a degradation of the nervous system. Of similar importance is also the loss of morpho-computational power by changes of the mechanical proper- ties of the patient’s body. If true, this hypothesis suggests novel types of support systems, one of which we presented in earlier works. The second hypothesis regards models that describe the dynamics of a population of tumor cells. We discuss a way to optimize synergistic treatments (combination of hyperthermia and radiotherapy) that is based on the assumption of a low-dimensional control system. Finally, the third hypothesis presents a morphological implementation of Matzinger’s mechanism of danger signals in immunology. The model we present is inspired by studies in the origin-of-life research, where possibilities to control molecular parasitism are discussed. The article closes with the discussion of a scientific roadmap for the application of morphological control in the field of medicine.
Authors: Rudolf M. Füchslin, Helmut Hauser, Irene Poli, Roberto Serra, Marco Vilani, Stephan Scheidegger, and Mathias S. Weyland
Chapter 19: The Morphological Computation Principles as a New Paradigm for Robotic Design
Abstract: A theory, by definition, is a generalization of some phenomenon observations, and a principle is a law or a rule that should be followed as a guideline. Their formalization is a creative process, which faces specific and attested steps. The following sections reproduce this logical flow by expressing the principle of Morphological Computation as a timeline: firstly the observations of this phenomenon in Nature has been reported in relation with some recent theories, afterward it has been linked with the current applications in artificial systems and finally the further applications, challenges and objectives will project this principle into future scenarios.
Authors: Davide Zambrano, Matteo Cianchetti, and Cecilia Laschi
Chapter 20: Morphological Computation - The Body as a Computational Resource
Abstract: Recently, two theoretical models for morphological computation have been proposed [13, 14]. Based on a rigorous mathematical framework and simulations it has been demonstrated that compliant, complex physical bodies can be effectively employed as computational resources. Even more recent work showed that these models are not only of theoretical nature, but are also applicable to a number of different soft robotic platforms. Motivated by these encouraging results we dis- cuss a number of remarkable implications when real physical bodies are employed as computational resources.
Authors: Helmut Hauser, Kohei Nakajima, and Rudolf M. Füchslin