Design and Improvement of Brain-Machine Interface Using Portable NIRS Device

近年来，机器人技术和医学科学一直在关注一种新的输入接口，即“脑机接口”。如果一个实用的BMI系统得以实现，它将在许多新领域具有开创性的应用，例如人机交互和特殊需要人士的支持性护理。因此，我们的目标是使用便携式近红外光谱（NIRS）设备设计一个实用的BMI系统。在这项研究中，我们开发了一种基于lvq的大脑状态分类器，用于便携式近红外光谱仪设备，该设备能够对几种人类心理图像（例如，给予机器人的心理命令）和人类情感对应的脑血流模式进行分类。然后，讨论了基于nir的BMI系统在日常环境中的应用问题，并设计了改进的基于lvq的分类器。除此之外，我们还通过实验研究了脑血流对外部视觉刺激的反应。

In recent years, robotics and medical sciences have focused on a new input interface, "the brain-machine interface." If a practical BMI system is realized, it could have pioneering applications in various new fields, e.g., human-robot interaction and supportive care for persons with special needs. Therefore, we aimed to design a practical BMI system using a portable Near-Infrared Spectroscopy (NIRS) device. In this research, we developed a LVQ-based cerebral state classifier for a portable NIRS device that enabled to classify cerebral blood flow patterns corresponding to several human mental imagery (e.g, mental commands given to a robot) and human emotions. Then, we discussed problems to utilize the NIRS-based BMI system in the daily environment and designed the improved LVQ-based classifier. In addition to that, we experimentally investigated the response of the cerebral blood flow against external visual stimuli.