A study on Therapy Robot to estimate human emotion based on touching motion

这项研究的目的是通过人类接触宠物类机器人的动作来估计情绪。在机器人上实现了由多个弯曲传感器和电容传感器组成的触摸传感器来测量手的触摸运动。利用改进的基于自组织映射或感知器学习算法的情感估计技术，将人体的碰撞、推搡、抚摸和颤抖等触摸动作识别并呈现在二维地图上。

This research aims to estimate emotion through the motion of the human that touches a pet type robot. A touch sensor combined with several bending sensors and capacitance sensors is implemented on the robot to measure the touching motion of the hands. By using the developed emotion estimating technique using Self-Organization Map or perceptron learning algorithm, human touching motion such as hitting, pushing, caressing and trembling, is distinguished and presented on a two dimensional map .