Restaurant Environment Customer’s Lifelong Incremental Action Sequence Learning Self-Motivated Developmental Cognitive Robot

Beom-Jin Lee, Chang-Min Choi, and Byoung-Tak Zhang


In the field of cognitive science, robotics has been widely used to experiment cognitive behaviors. As an extension, the main idea in this paper is based on the infant’s developmental process which enables goal directed self-exploration and social interaction skills. Inspired by the studies of cognitive developmental robotics, we designed a model which can self-organize and explore the optimal action generation based on human customer’s action in restaurant environment. Beginning with self-organization process of humanoid robot, we expect the agent to automatically find the movement to achieve simple action movement. Furthermore, applying Lifelong Incremental learning, the process increases the skills of the agent to learn various situation dependent skills which could obtain highest reward for the given situation. An experiment has been modified and the result shows that after the trial and error repetition, the agent enables to perform diverse situated actions.