基于改进EMD消除激光多普勒信号中的趋势项  

A Trend Term in the Elimination of Laser Doppler Signal Based on Improved EMD

作　　者：范国浩 李玲玲 张艳兵[3] 马铁华[3] FAN Guohao;LI Lingling;ZHANG Yanbing;MA Tiehua(National Computer Systems Engineering Research Institute of China,Beijing 100083,China;No.6904 Factory of PLA,Taiyuan 030051,China;National Key Laboratory for Electronic Measurement Technology,North University of China,Taiyuan 030051,China)

机构地区：[1]华北计算机系统工程研究所,北京100083 [2]解放军第6904工厂,太原030051 [3]中北大学电子测试技术国家重点实验室,太原030051

出　　处：《弹箭与制导学报》2017年第6期153-156,164,共5页Journal of Projectiles,Rockets,Missiles and Guidance

摘　　要：在使用自由式微小型Hopkinson杆的窄脉冲校准系统对高g值加速度传感器进行校准时,激光多普勒信号含有的趋势项会对之后的解算造成影响,而采用EMD分解法消除信号趋势项时会产生端点效应。针对此提出了基于最小二乘支持向量回归机的EMD分解法。通过算法仿真以及对实测多普勒信号趋势项的提取,实验结果表明:使用改进后的EMD分解法对趋势项进行提取可有效消减端点效应,效果明显,提高瞬时相位解算的精度。When using the narrow pulse calibration system of micro Hopkinson bar to calibrate the high-g acceleration sensor,the trend item contained in laser Doppler signal would impact the subsequent operation.When EMD decomposition was used to eliminate the signal trend,the end effect would occur.Focusing on these problems,this paper presents an EMD decomposition method based on least square support vector regression machine.Through the simulation and the extraction of the measured Doppler signal trend,the experimental results show that using the improved EMD decomposition method to extract trend items can effectively reduce the end effect,and the effect is obvious.It also can improve the accuracy of instantaneous phase.

关 键 词：多普勒信号 趋势项 最小二乘支持向量回归机 改进EMD 

分 类 号：TN249[电子电信—物理电子学]