Abstract
In this paper, we propose a learning model based on psychology for a moving robot to give robot like creature.The proposed model voluntarily acts to satisfy the robot's desire on the basis of operant conditioning in behavioral psychology.By incorporating this model into the robot, we expect that the robot can provide the robot with a sense of affinity.The robot's action is generated by a recurrent neural network (RNN).Therefore the robot learns the action in consideration of time series.In this paper, we run simulation about a move action learning model to verify that the robot moves like creature.We confirm that the model indicates that the robot learns the move actions which correspond to the robot's desire, and the robot has behavior of the operant conditioning.
