Autonomous action selection with motivation-based consciousness and behavior architecture of animal

Abstract

This paper presents an autonomous action selection system that incorporated a motivation model, and a motion control system to achieve natural behaviors showing human emotion such love, pity etc. We have attempted to give a robot "consciousness" and "emotion" such as that identified in humans and animals in order to enhance the affinity between humans and robots. A hierarchical structure model for such a robot has been developed to connect the robot's consciousness with the robot's behavior. However, it is difficult to autonomously control the timing and the motion that selects the consciousness and behavior of the robot, whatever the time and position. A motivation model has therefore been developed, and it was combined with the hierarchical structure model to induce and autonomously select consciousness and behavior. The action of dopamine in neurotransmitters was then incorporated into the motivation model to add activity to the robot in conjunction with the incentive to perform a behavior. As a result, this system allows a robot to behave naturally in any situation.