Using Hormonal Feedback to Modulate Action Selection in a Competitive Scenario

In this paper we investigate the use of hormonal feedback as a mechanism to modulate a "motivation-based," homeostatic action selection mechanism (ASM) in a robot. We have framed our study in the context of a dynamic, multirobot, competitive "two-resource" action selection problem. The introduction of competitors has important consequences for action selection. We first show how the interaction between robots introduces new forms of environmental complexity that affect their viability. Secondly, we propose a "hormone-like" mechanism that, modulating the input of the ASM, tackles these new sources of complexity.