ABSTRACT. Fireflies communicate by producing bioluminescence to signal their presence and court mates. In particular, some species emit patterns of short flashes that have the potential to encode information; males flash according to a species-specific pattern in order to attract and locate females. As multiple firefly species can share the same habitat, potential visual clutter could greatly hinder species discrimination. We investigate how flash sequences can co-evolve to be distinguishable by developing a method for simulating sequences that minimize similarity and energetic cost. We observe an emergent periodicity in the resulting optimal sequences despite the lack of any constraints on the sequences to be periodic. We also demonstrate a method of reconstructing potential cost functions from the phylogenetic relationships of extant species alongside their characteristic flash patterns.

ABSTRACT. In this paper, we introduce a recovery approach to deal with the joint failure problem in snake robots. As the failed joint breaks up the overall shape of the robot, in the proposed method, the desired robot shape is restored by utilizing the hyper-redundancy of the snake robot structure. We control the adjacent joints of the failed joint to keep the robot shape close to its original form before failure. Free and locked failures are considered in our approach. The validity of the recovery method was verified in the simulation environment, considering various locations of the failed joint. The results showed that the robot was able to resume the locomotion by retaining its desired shape despite having a failed joint.

ABSTRACT. In recent years, extendable snake-like robots are attracting researchers' interests. A large number of degrees of freedom makes it possible for the robots to perform new gaits. It is expected that the telescopic motion can also give the robots higher adaptability for the environment. In this paper, we propose a controller of an extendable snake-like robot for head trajectory tracking. The validity of the controller is tested via physical simulations.

ABSTRACT. Snakes possess versatile gait patterns and use them appropriately to adapt to various environments. To reproduce this ability, we have previously proposed a decentralized control scheme for snake robots based on {¥it Tegotae}, a Japanese concept describing how well a perceived reaction matches an expectation. In this paper, we briefly review the proposed control scheme and the simulation results. We also explain its limitation and discuss a possible solution to it.

ABSTRACT. Quadrupeds have asymmetric body structure, which gives the offset of the center of mass (CoM) in the anteroposterior direction. Horses, which have long necks, don't use rotary gallop but transverse gallop when traveling at their high-speed range. Since the CoM offset affects the running dynamics, we expected that it also affects the gait selection by horses. We analyzed the bounding gait using a dynamical model with a fore-hind offset of CoM. We obtained periodic solutions of the model by numerical calculations and classified them based on their sequences, and examined their stability, maximum ground reaction force, and forward transport efficiency. The result revealed the sequence corresponding to the transverse gallop was optimal for horses, suggesting the reason for the gait selection of horses.

ABSTRACT. Towards generating stable and efficient locomotion on slippery ground, several indirectly controlled sliding robots have been designed in the previous works. Although numerical simulations showed the possibility of achieving high-speed forward sliding locomotion, it is difficult to realize similar performance with a real machine due to the unexpected behavior of their telescopic wobbling motion. In addition, it is also inefficient to conduct an optimization process for them due to their redundant parameters. In this work, we introduce a minimalistic model for a sliding locomotion robot. The telescopic wobbling motion is replaced by rotating motion while remaining strong propulsive force. Equations of motion are analytically derived and the validity is numerically proved.

ABSTRACT. Passive dynamic walking on a gentle slope is the most ex- cellent gait of legged robots in terms of its energy efficiency, since all of the energy required to make a forward stepping is supplied by gravity only. To extend such gait to level ground, or more challenging condition, e.g., rough terrain, a fundamental task is to recycle the energy consumed by ground collision. Inspired by soft tissues of animal and human, wob- bling masses are attached to the walkers in the previous studies. Here we investigate the effect of different kinds of wobbling masses on the gait performance. First, a biped walker with a 2-DOF passive wobbling mass model is derived. Second, passive dynamics walking gait of this robot is numerically generated. Finally, the motion analysis is conducted under different wobbling conditions. Our results shows that the walking speed can be improved by up-and-down oscillation, however, worsened by the rotational wobbling.

ABSTRACT. Cheetahs run at remarkably high speed involving the large spine movement. In this paper, to clarify the effects of the spine movement during galloping, we developed a simple model with body flexibility and find periodic solutions by using parameters from measured data. We numerically searched periodic solutions and compared them based on the evaluation indexes for the ground reaction force and energy efficiency. The results revealed that the spine movement improves the running performance not only in a kinematic but also in a dynamic perspective. From these results, the mechanism under which cheetahs achieve high-speed running with the spine movement is suggested.

ABSTRACT. In this workshop, we will discuss the identities of natural and artificial swarm agents and explore how embodiment can contribute to the perpetuation of deployed multi-agent systems. We invite contributions from the diverse DARS-SWARM2021 community, representing the intersection of biology and engineering, to feed the discussion on what makes an embodied swarm agent and what swarm agents of the future will look like. Featured research may include but is not limited to swarm robots, living agents, chemical systems and embodied swarm algorithms.

ABSTRACT. In nature, a group of organisms is formed by increasing the number of the organism through the division process. The characteristics of the resulting group of organisms are organized through the division process, and the behaviors of the group affect each individual and the division process. In this study, we aim to investigate this mutual influence between collective behavior and the division process. For this purpose, we recorded the growing population of the ciliate, Tetrahymena from one individual for several generations and analyzed the motion by tracking each individual. For comparison, we also recorded the motion of the population of Tetrahymena not divided from one individual. We found that the pattern of motion was inherited, but also the overall dynamics changed over the generations. We discuss how phenotype in the collective behavior is organized and inherited in the group of Tetrahymena.

ABSTRACT. Machine learning algorithms are becoming an important tool in the field of biology/ethological analysis. Here, using unsupervised learning techniques and massive dataset of the honeybee's locomotions in a hive, we analyze the time series of individual behaviors (x,y position, kinetic energy, and roles such as foraging, dancing) and discuss the development process of the hive. The goal of this project is to examine the collective behavior and the division of labor of honeybees from a new aspect, independent of the biases of human observation.

ABSTRACT. This timely project, that explores social isolation, digital cultural participation, and the integration of embodied agents within social ecosystems, will be focused on the development of a prototype kit of deployable robots in response to the question: “‘How can I be present and connect with other bodies in a space where I am not?”. By tele-operating robot avatars from their homes, via video call, people experiencing isolation and loneliness will be given the capability to utilize embodied agents as a means to participate in arts and cultural events. We will work with a group of socially isolated older and younger people to co-design a swarm of miniature, soft robots through a series of workshops. The work will exploit the inherent safety of soft-bodied robots for human interaction, and take soft robotics in a novel direction by using participatory research as a tool for robot morphological design. 

ABSTRACT. As tiny robots become individually more sophisticated, and larger robots easier to mass produce, a breakdown of conventional disciplinary silos is enabling swarm engineering to be adopted across scales and applications, from nanomedicine to treat cancer, to cm-sized robots for large-scale environmental monitoring or intralogistics. This convergence of capabilities is facilitating the transfer of lessons learned from one scale to the other. Cm-sized robots that work in the 1000s may operate in a way similar to reaction-diffusion systems at the nanoscale, while sophisticated microrobots may have individual capabilities that allow them to achieve swarm behaviour reminiscent of larger robots with memory, computation, and communication. Although the physics of these systems are fundamentally different, much of their emergent swarm behaviours can be abstracted to their ability to move and react to their local environment. This presents an opportunity to build a unified framework for the engineering of swarms across scales that makes use of machine learning to automatically discover suitable agent designs and behaviours, digital twins to seamlessly move between the digital and physical world, and user studies to explore how to make swarms safe and trustworthy. Such a framework would push the envelope of swarm capabilities, towards making swarms for people.

ABSTRACT. In this talk, the three related topics of robotic self-reconfiguration, self-repair, and self-replication are discussed. This will include a review of past works by many authors, and future directions. Current work on multi-robot team diagnosis and information fusion will also be discussed. This leads to probabilistic formulations of the health of a robotic team, involving sensor uncertainties and calibration issues. In order to quantify the robustness of self-replicating robots, measures of the degree of environmental uncertainty that they can handle need to be computed. The entropy of the set of all possible arrangements (or configurations) of spare parts in the environment is one example of such a measure and has led us to study problems at the foundations of statistical mechanics and information theory. The use of robots to harvest resources in outer space to the benefit of humanity will require robust autonomous teams that can handle uncertainty, and function reliably while using in situ resources to repair and reproduce.

ABSTRACT. Our research team has been studying feral horses living in northern Portugal since 2015. The population of horses at this research site consists of up to 230 individuals belonging to over 30 different unit groups. We use drones, unmanned aerial vehicles, to observe their behavior from the sky. The horses live in a relatively plain grassland environment and there are few obstacles between the drones in the sky and the horses on the ground, thus it is possible to identify and record all of the individuals in clear visibility. In this talk I will summarize our findings. First, spatial relationships within a unit group of horses showed that the stallion male was most commonly located in the periphery of his group, while females were located in the center. Second, the distance between two horses peaked at around 4 meters, suggesting that a repulsion and attraction force exists between individuals. Third, a mathematical simulation that incorporated attraction and repulsion forces was able to explain our empirical data of inter-individual spatial distributions. Fourth, analyses of spatial relationships at an inter-group level showed the existence of a higher level of social structure, suggesting that feral horses form a multi-level society. Our findings shed new light on the evolution of sociality in the mammalian lineage.