Mondada, FrancescoMills, Robert MatthewBotner Barmak, Rafael2024-07-022024-07-022024-07-02202410.5075/epfl-thesis-10594https://infoscience.epfl.ch/handle/20.500.14299/208941Social insects, such as ants, termites, and honeybees, have evolved sophisticated societies where the collaborative efforts of "simple" individuals can lead to the emergence of complex dynamics. The reliance of each organism on the collective is so great that individual organisms would not survive outside the group; for this reason, such species are considered "superorganisms". Studying superorganisms opens a unique window into collective cognition and the emergence of self-organized behaviors, where entire colonies composed of individuals with limited cognitive abilities (ie, small brains) and the absence of a leader to orchestrate actions are able to reach consensus and coordinate actions collectively. However, investigating such animal societies is not trivial. Most social insects live in large, cryptic colonies and can exhibit aggressive behaviors. Consequently, experiments are often conducted with a limited number of organisms, usually for short periods of time, and under artificial conditions that incompletely reflect their natural habitat. To address these challenges, we employed a novel methodology to guide the development of a robotic system capable of forming a cohesive mixed society with honeybees (ie, a biohybrid superorganism) that would allow long-term investigations of socially complete colonies under natural conditions. We focused on the western honeybee (Apis mellifera), not only for its agricultural importance as an efficient and domesticated pollinator but also for its shared characteristics and behaviors with other eusocial insects. The primary objectives of this research were to design a biocompatible robot and examine the honeybee society's collective thermoregulatory behaviors, essential for maintaining colony homeostasis during the foraging and "hibernation" seasons. Here, we detail the design, development, and validation of the robotic device aimed at investigating collective thermoregulatory behaviors through arrays of thermal sensors and actuators. During a series of experiments, in which robotic devices were installed into beehives populated with thousands of bees, we demonstrated the robots' acceptance and ability to interact with host colonies. During foraging and overwintering periods, the robotic system was able to observe the broodnest thermoregulation and winter cluster thermogenesis patterns over 121 days. Furthermore, via two perturbation experiments, we showed that the robotic device reliably modulated the superorganism's spatiotemporal reorganization in response to thermal stimulation. In one experiment, a single robotic frame lured a winter cluster to perform a repeating zigzag motion that lasted 51 days. In a second experiment, two robotic devices orchestrated the thermal stimuli, attracting bees to perform a 2D path between the surface of the two frames for 35 days. Additionally, upon identifying the thermal collapse of a colony, we used the robotic system in a "life-support" mode through its thermal actuators. Finally, we demonstrated a closed-loop interaction with a cluster of thousands of bees, with the robotic system autonomously deciding when to intervene based on the perceived colony state. These results demonstrate the potential of biohybrid societies to provide insights into the types of information that colonies perceive as relevant, thereby confirming or challenging our current understanding of self-organization and collective behaviors in complex animal societies.enanimal-robot interactioninteractive roboticsethoroboticsbiohybrid systemsmixed societiescollective behaviorsthermoregulatory behaviorssocial insectshoneybeesthesis::doctoral thesis