ABSTRACT. In this talk I will briefly survey some of the results we obtained in the last ten years with the goal of illustrating the main lessons learned. I will then focus on two current research directions that try to address some of the problems we identified in our previous research: mergeable nervous system, where we study how self-organisation can be made more powerful as a tool to coordinate the activities of a robot swarm by injecting some components of hierarchical control; and robustness to malicious robots via blockchain-based smart contracts, that is, how to control robots in a swarm using smart contracts so that their are resistant to malicious robots and to Sybill attacks.

ABSTRACT. Swarms of insects, schools of fish and flocks of birds display an impressive variety of collective movement patterns that emerge from local interactions among group members. These puzzling phenomena raise a variety of questions about the interactions rules that govern the coordination of individuals’ motions and the emergence of large-scale patterns. While numerous models have been proposed, there is still a strong need for detailed experimental studies to foster the biological understanding of such collective motion phenomena. I will first describe the methods that we have developed in the recent years to characterize social interactions between individuals involved in the coordination of swimming in Rummy-nose tetra (Hemigrammus rhodostomus) from data gathered at the individual scale. This species of tropical fish performs burst-and-coast swimming behavior that consists of sudden heading changes combined with brief accelerations followed by quasi-passive, straight decelerations. Our results show that both attraction and alignment behaviors control the reaction of fish to a neighbor. Then I will present how these results can be used to build a model of spontaneous burst-and-coast swimming and social interactions of fish, with all parameters being estimated or directly measured from experiments. This model shows that the simple addition of the pairwise interactions with two neighbors quantitatively reproduces the collective behavior observed in groups of five fish. Increasing the number of interacting neighbors does not significantly improve the simulation results. Remarkably, we find that groups remain cohesive and polarized even when each agent only interacts with only one of its neighbors: the one that has the strongest contribution to the heading variation of the focal agent. Finally, I will present a swarm robotic platform with which we investigate the impact of collision avoidance protocol based on speed control on the group behavior. This platform combines the implementation of the fish behavioral model and an engineering-minded control system to deal with real-world physical constraints. Remarkably, and as already observed in the model simulations, even when robots only interact with their most influential neighbor, our results show that the group remains highly cohesive and polarized while reproducing the behavioral patterns observed in groups of five fish. Overall, our results suggest that fish have to acquire only a minimal amount of information about their environment to coordinate their movements when swimming in groups.

ABSTRACT. This research focuses on the mathematical modeling aspect of self-propelled particles. Several interesting collective motion patterns are seen in self-propelled particle systems such as schooling fish. The milling state is one of those characteristic patterns that can be captured, for instance, using the Rayleigh-Morse equations. This model only considers the direct interactions between individual fish, neglecting the interactions between fish and the surrounding environments such as walls and obstacles. Different shapes of walls and obstacles in the calculation domain may trigger transitions from one metastable state to another, and may influence the final form of emerging patterns and their converging behaviors. We will simulate and investigate those effects to the collective motion patterns, envisaging several possible scenarios. Furthermore, for understanding more complex behaviors interacting with the surrounding environments, we believe that we need to extend the mathematical model by considering the effect of water. This extension ends up with an extended version of Rayleigh-Morse equations. We will demonstrate the effectiveness of the extended model through numerical examples. Finally we will discuss the possibilities of further developments of this research.

ABSTRACT. Controlling and optimizing pedestrian flows are critical in urban planning as walking is the most common form of mobility among increasingly diverse transportation services. Here we investigate pedestrian counter flows in a straight corridor, in which two groups of people are walking in the opposite directions, by means of a molecular dynamics approach with the social force model. We demonstrate that a simple array of obstacles improves flow rate even in a crowded situation as a result of flow separations, that is, the obstacles separate groups of pedestrians waking in the opposite directions. The appropriately designed obstacles are fully capable of controlling the filtering direction, such that pedestrians are tend to keep left spontaneously. The present results potentially provide a guideline for industrial design to improve daily human mobility.

ABSTRACT. Systems with many individuals such as flocks, swarms or herds of animals typically show a parameter dependent macroscopic, i.e. low-dimensional coherent behavior which is of interest to be analyzed. This analysis is in particular difficult if the underlying microscopic system is either given as discrete agent-based model or as laboratory experiment. A mathematical method will be presented which allow to perform a numerical bifurcation analysis of the macroscopic behavior for a given microscopic model or even an experiment. Examples will be given mainly from pedestrian flows. The method allows to visualize and investigate unstable pedestrian flows.

ABSTRACT. A new approach for solving combinatorial optimization problem incorporating a traditional mathematical problem of chases and escapes is proposed. In addition to the steepest descent and neighboring search, It performs a chase-and-escape game on the ``landscape'' of the cost function along with the steepest descent. We evaluated the algorithm for the Traveling Salesman Problem, which preliminarily show that this new approach can be fruitful.

ABSTRACT. Topology optimization is a general technique of structure designs, which has broad range of applications. While traditional studies in topology optimization mainly focused on centralized computation of a static structure, we focus on a problem of decentralized optimization where a swarm of multi agents self-organizes a structure dynamically with their ability to move and connect to each other. We propose a distributed optimization algorithm of dynamic structure design as an extension of Evolutionary Structural Optimization method with the game theoretic foundation.

ABSTRACT. The effect of having heterogeneity in a group is studied in a simple model of group chase and escape problem. We introduce lazy chasers in two ways: uniformly and in a “division of labor” way. It is shown that while the former is always ineffective, the latter can improve the efficiency of catching, through the formation of pincer attack configuration by diligent and lazy chasers.

ABSTRACT. Many small organisms such as bacteria can attract each other by depositing chemical attractants. At the same time, they exert repulsive force on each other when crowded, which can be modeled by effective pressure as an increasing function of the organisms' density. As the chemical attraction becomes strong compared to the effective pressure, the system will undergo a phase transition from homogeneous distribution to aggregation. In this work, we describe the interplay of organisms and chemicals on a two-dimensional disk with a set of partial differential equations of the Patlak-Keller-Segel type. By analyzing its Lyapunov functional, we show that the aggregation transition occurs discontinuously, forming an aggregate near the boundary of the disk. The result can be interpreted within a thermodynamic framework by identifying the Lyapunov functional with free energy.

ABSTRACT. This paper focuses on the effect of collisions among robots on the collective behavior in robotic swarms. The research field of swarm robotics emphasizes the importance of the embodiment of robots; however, only a few studies have discussed how it influences the collective behavior of robotic swarms. In this paper, the path formation task of robotic swarms was conducted in computer simulations with and without considering robot collisions. Additionally, the experiments were performed with varying the size of robots. The robot controllers were obtained by an evolutionary robotics approach. The results show that the robot collisions would affect not only the performance of the robotic swarm but also the specialization among robots to manage congestion. The robot collisions seem to provide negative feedback on robotic swarms to emerge the specialization.

ABSTRACT. We are interested in programming a swarm of molecular robots that can perform self-assembly to form specific shapes at a specific location. Programming such robot swarms is challenging for two reasons. First, the goal is to optimize both the parameters and the structure of chemical reaction networks. Second, the search space is both high-dimensional and deceptive. In this paper, we show that MAP-Elites, an algorithm that searches for both high-performing and diverse solutions, outperforms previous state-of-the-art optimization methods.

ABSTRACT. Artificial life is the study of artificial systems that exhibit the behavioural characteristics of natural living systems. One of the aims of artificial life research is to synthesize artificial cells from chemical precursors. The challenge of creating an artificial cell composed of all attributes of living counterparts (such as growth and development, homeostasis, movement, energy use) is one that is yet to be solved. Our ambition is to create a system that will at least partly mimic cell behaviour. We focus on the investigation of organic droplets in the presence of aqueous solutions of surfactants. We have found that in a similar way to living cells, these droplets are able to perform chemotactic movements in concentration gradients of chemoattractants, change their shape or behave collectively. Recently we proposed to call such droplets with life-like behaviour “liquid robots”. Droplets defined as liquid robots are a novel concept that could find engineering applications in various areas, such as environmental sensing and remediation.

ABSTRACT. The collective migration of thousands of living cells is key to many biological processes. The ability to herd such cellular ensembles would offer exciting possibilities in biology and bioengineering. Previously, we demonstrated uniaxial control of collective cell migration using a bioelectric phenomenon known as electrotaxis. Here, we have developed a new device incorporating multiple addressable electrodes under computer control that allows us to literally program collective cell migration in real-time. We present preliminary results on such herding in living skin samples.

ABSTRACT. In our body, cells of epithelial tissues pack into a homeostatic state in which cellular dynamics show striking similarity to dynamic heterogeneity analogous to dense granular matter or super-cooled liquid. Although several studies in the past have tried to elucidate the nature of this heterogeneity using epithelial cell monolayer, its physiological relevance remain elusive. Here we explore the role of dynamic heterogeneity in Madin-Darby Canine Kidney (MDCK) epithelial cells during wound closure, in which these cells are required to migrate in a collective and coordinated fashion towards the denuded area. Formation of leader cells during such a cellular migration event is previously shown by several studies, however, how cells select their leaders from a seemingly uniform epithelium remain unclear. We report that owing to the dynamic-heterogeneity of the monolayer, cells behind the prospective leaders manifest locally increased traction and monolayer stresses much before these leaders display any phenotypic traits. Followers, in turn, pull on the future leaders to elect them to their fate. Interestingly, frequency at which leader cells emerge at the wound margin depend upon the length scale of force correlation within the monolayer. In fact, by modifying force transmitting ability of cells, we were able to modify the number of leader cells at the wound margin. Once formed, these leaders pull the followers towards the direction of migration and the territory of a leader cell extend to the length up-to which leader cells can transmit forces which, strikingly, is similar to the distance up-to which force is correlated within the monolayer, bringing us back to the dynamic heterogeneity of the epithelium. In the end, we attempted to vary the dynamic heterogeneity itself, by changing cell density, and found that this also altered the number of leader cells at the wound margin. Interestingly, at low heterogeneity, in spite of the large number of cells showing leader-like morphology, wound closure was slower, due to decreased persistence, decreased coordination, and disruptive leader-follower interaction. Together, these results demonstrate the physiological relevance and importance of dynamic heterogeneity in epithelial tissues, especially with respect to the hierarchies of cell collectives during wound healing.