激光多普勒技术的机电系统运动模式识别  

Motion pattern recognition of electro-mechanical system based on laser Doppler technology

作　　者：沈红伟 刚建华 SHEN Hongwei;GANG Jianhua(Department of Mechanical and Electrical Engineering,Cangzhou Normal University,Cangzhou Hebei 061001,China)

机构地区：[1]沧州师范学院机械与电气工程学院,河北沧州061001

出　　处：《激光杂志》2020年第11期115-119,共5页Laser Journal

基　　金：河北省科技厅项目(No.18211844)。

摘　　要：机电系统体积不大、运动速度快,但运动特征易被噪声所影响,导致以往采用光学检测法识别机电系统运动模式时,检测信号存在误差,运动模式识别精度较低。提出基于激光多普勒技术的机电系统运动模式识别方法,通过基于激光多普勒的机电运动特征检测方法,获取机电系统运动特征,并在检测过程中引入自适应滤波器,去除噪声信号的干扰,提高特征检测精度;再基于获取的运动特征,采用基于Relief-F的SVM运动模式分类算法实现机电系统运动模式识别。实验结果表明,方法对机电系统里谐振器的运动模式识别精度高,在低噪与高噪条件下,可实现抗干扰、高精度的机电系统运动模式识别。The mechanical and electrical system is small in size and fast in motion,but its motion characteristics are easily affected by noise.As a result,when optical detection method is used to identify the motion mode of mechanical and electrical system in the past,the detection signal has errors and the recognition accuracy of motion mode is low.In this paper,a method of motion pattern recognition of electro-mechanical system based on laser Doppler technology is proposed.Through the detection method of electro-mechanical motion feature based on laser Doppler,the motion feature of electro-mechanical system is acquired,and an adaptive filter is introduced in detection process to remove the interference of noise signal and improve the accuracy of feature detection.Then,based on acquired motion feature,the SVM motion mode based on the Relief-F adopted classification,the algorithm is used to realize the motion pattern recognition of electro-mechanical system.The experimental results show that the accuracy of proposed method is high,and under the condition of low noise and high noise,the anti-interference and high-precision electro-mechanical system motion pattern recognition can be realized.

关 键 词：激光 多普勒 机电系统 运动 模式 识别 

分 类 号：TN958.98[电子电信—信号与信息处理]