Behavior Selection Method of Robots Based on a Markovian Emotional Model

Abstract

In this paper, we propose a new decision making system of autonomous mobile robots based on a Markovian emotional model. In the proposed method, it is possible to build a method of determining behavior along the task by optimizing system parameters automatically. In addition, we propose a method for self-organization of a nonlinear transformation that makes it possible to convert input stimuli to internal parameters properly using a neural network. The effectiveness of the proposed method is verified through simulations of identifying unknown environments by evaluating the exploration time under five different test environments. These results showed that the automatic construction of the decision making system was performed, and valid action decisions were made effectively.