Open-end human robot interaction from the dynamical systems perspective: Mutual adaptation and incremental learning

This paper describes interactive learning between human subjects and robot using the dynamical systems approach. Our research concentrated on the navigation system of a humanoid robot and human subjects whose eyes were covered. We used the recurrent neural network (RNN) for the robot control. We used a "consolidation-learning algorithm" as a model of hippocampus in brain. In this method, the RNN was trained by both a new data and the rehearsal outputs of the RNN, not to damage the contents of current memory. The proposed method enabled the robot to improve the performance even when learning continued for a long time (open-end). The dynamical systems analysis of RNNs supports these differences.