SSVEP-based hierarchical architecture for control of a humanoid robot with mind

In this paper, we present an SSVEP-based hierarchical architecture for control of a humanoid robot with mind, consisting of five-layer decisions from robot state level to implementation level. This architecture is able to control a variety of humanoid robot behaviors at different levels to perform a complex task. We implement this hierarchical architecture on our Cerebot platform and test this control architecture in a multi-task experiment. We compare its control performance with the one achieved using manual control by an experienced operator. The results show that the architecture with coordinating multi-layer decisions and fusing human and robot actions is a good way to solve the issue of the limited information transfer rate (ITR) in mind-controlled humanoid robot system using current EEG-based BCI technology.