2P2-L11 Design of Brain-machine Interface Using Near-infrared Spectroscopy : Study on learning condition for improvement of classification performance(Neurorobotics & Cognitive Robotics)

摘要

Recently, brain-machine interface (BMI) systems is focused in the robotics and medical sciences. In this research we discuss a design problem of a BMI system using near-infrared spectroscopy(NIRS) and developed LVQ-based classifier for several patterns of cerebral blood flow corresponding to human physical motions, human mental imagery, mental commands given to a robot and human emotions. In this paper, we discuss a suitable learning condition of the prepared classification.