Theoretical Construction of an Emotion-Linked Behavior Formula and Its Application in Humanoid Robotics

Expansion Based on Simonov's Emotion Theory and Application in Emotional Behavior Regulation of Humanoid Robots

Abstract

At present, there exist numerous theoretical formulations regarding emotions. However, once a behavior initiated by a certain emotion is put into practice, the execution of such behavior will introduce additional stimuli, which in turn generate new emotions. In this paper, these influencing factors are comprehensively considered. On the basis of drawing on Simonov's emotional theory, the algorithms of necessary information and available information are explained in detail, and a behavioral formula is constructed. This behavioral formula mainly consists of a main-process term and a sub-process term. The sub-process term is triggered by the main-process term and exerts either a promoting or inhibiting effect on the latter. The formula is then illustrated with common examples, including decision-makers who do not participate in behavioral output, individuals who claim a share of the benefits but refuse to engage in any affairs, independent entrepreneurs, wage earners, group members who abandon themselves to degeneration, and those who exert their utmost efforts for the group. This illustration also serves as a verification of the formula’s validity. Meanwhile, combined with generative large language models (LLMs) and the Cognitive Process and Information Processing Model (CPIPM), the information generation method in the formula is elaborated, aiming to provide valuable insights for the development of humanoid robots.