基于表面肌电信号的虚拟座舱动作模式识别  

Pattern Recognition of Movement Based on sEMG in Virtual Cockpit Environment

作　　者：刘礼浩 吴东苏[1] 金亮[1] 吴文丽[1] 

机构地区：[1]南京航空航天大学民航学院飞行模拟与先进培训工程技术研究中心,江苏南京211106

出　　处：《航空计算技术》2015年第5期28-32,共5页Aeronautical Computing Technique

基　　金：国家自然科学基金项目资助(51205195;61039002);江苏省自然科学基金项目资助(BK20130981);南京航空航天大学研究生创新基金项目资助(KFJJ20130223;KFJJ201472)

摘　　要：在虚拟座舱交互头部运动中,长时间使用虚拟现实头盔时仿真参与者会出现了失去方向、恶心呕吐等仿真病。其中,仿真病产生的最重要的原因是视觉延迟。针对突然的头部运动产生的视觉延迟问题,提出了一种基于表面肌电(s EMG)信号的头部动作模式识别的算法。运用算法可成功地从胸锁乳突肌和头夹肌中识别出上仰、旋转和侧弯3种头部运动动作。实验结果表明,径向基函数神经网络具有识别能力强和学习速度快等优点,在模式识别领域具有广阔的发展前景。Simulation participants will lose the direction and appear the simulation of disease such as nausea and vomiting when they wear virtual reality helmet for a long time in virtual cockpit interaction head movement environment.Visual lag is the most important cause which generates the simulation disease.In view of visual lag problem caused by sudden rapid head movement,this paper presents a pattern recognition algorithm of head movement based on surface EMG signals .Applying this algorithm can successfully identify three types of movements of human head：nose-up pitch,rotating and lateral bending from cleidomastoid and musculi splenius capitis.The experimental result shows that RBF neural network has a great potential development in the field of pattern recognition because of the advantages such as high pattern recognition ability,fast learning speed,etc.

关 键 词：虚拟座舱 表面肌电信号 特征提取 径向基函数神经网络 模式识别 

分 类 号：TP241.3[自动化与计算机技术—检测技术与自动化装置]