基于小波包的人机协作训练生理心理特征提取与识别  

Psychophysiological feature extraction and recognition in human-robot cooperative training based on wavelet packet

作　　者：高翔[1] 徐国政[1] 郭猷敏[1] 梁志伟[1] 

机构地区：[1]南京邮电大学自动化学院,南京210023

出　　处：《仪器仪表学报》2012年第12期2875-2880,共6页Chinese Journal of Scientific Instrument

基　　金：国家自然科学基金(61104216);江苏省自然科学基金(BK2012832);江苏省高校自然科学基金(12KJB510015);南京邮电大学人才引进基金(NY211020;NY211067)资助项目

摘　　要：针对目前机器人辅助患者运动功能康复过程中人机协作训练机制主要停留在生物力学层次的局限,以机器人辅助患者远程运动康复为背景,研究了一种新的基于患者心理状态层次的人机生物协作训练方法。首先基于无线体域网技术构建一套远程无线智能康复监测系统,实时获取训练过程中患者自主神经系统生理参数;其次,运用小波包变换和线性分类器对采集到的多生理参数信息进行去噪、特征提取和心理状态识别;最后,选取健康受试者作为前期研究实验对象,结果显示训练过程中受试者心理状态平均识别率可达85%以上,进一步验证了系统的可靠性和算法的有效性。A new human-robot bio-cooperative rehabilitation training method for remote robot-aided motor neuroreha- bilitation is proposed based on the patient＇ s psychological state estimation, to solve the problem that human-robot co- operative training method in existing robot-aided motor functional rehabilitation system is still restricted to biomechan- ical efforts. Firstly, a remote wireless smart rehabilitation monitoring system using wireless body area network （WBAN） is constructed to collect the patient＇ s physiological parameters from autonomous nervous system, and then wavelet package transformation and linear discrimination analysis methods are used to extract and recognize the pa- tient＇ s physiological features and psychological state, respectively. Finally, some healthy subjects are recruited to perform psychophysiologieal experiment, and the results show that mean correct recognition rates during the course of rehabilitation training may be up to above 85% , which further verify the effectiveness of the system and the proposed algorithms.

关 键 词：远程康复 人机协作 小波变换 无线体域网 特征提取 

分 类 号：TP242.62[自动化与计算机技术—检测技术与自动化装置]