Towards an Anticipatory Mechanism for Complex Decisions in a Bio-Hybrid Beehive

Heinrich Mellmann, Volha Taliaronak, Verena V. Hafner

Abstract: To a certain extent, humans and many other biological agents are able to anticipate the consequences of their actions and adapt their decisions based on available information on current and future states of their environment. The same principle can be applied to enable decision-making in artificial agents. In order to decide on an action, an agent could envision the consequences for each of the actions and then choose the one promising the best outcome. This anticipatory scheme can enable fast decisions in highly dynamic and complex situations, which has been demonstrated in humanoid robots playing soccer. We extend this principle to the scenario of bio-hybrid beehives augmented with robotic actuators, which allow to influence the foraging locations of the bees. We investigate how a bio-hybrid beehive can make decisions and direct the bees in a way which would benefit the the whole ecosystem enabling sustainable beekeeping. We explore the general principles of anticipation and discuss connections to cognitive science and developmental robotics. We present an implementation of a simulator for the behavior of the augmented beehive and present preliminary results demonstrating the feasibility of the anticipatory approach.