An automatic control model for rat-robot

In this paper, a control model is developed to automate the process of navigation in rat-robot-a new type of bio-robot based on BCI (Brain-Computer Interface) technique. Because of the particular difficulties in rat-robot control, we design a novel control model to `learn' and `imitate' the control behavior of human guidance. General Regression Neural Network (GRNN) model is used to analyze the control commands made by human operators, with the locomotion information of rat-robot recorded and analyzed in a video-based experimental system. The results of the control model shows that the human control process could be well understood and predicted, and expected to generate control commands automatically in future real-time rat-robot navigation experiments.