摘要
针对输电信号在远程传输过程中受到环境噪声扰动影响信号识别的问题,本研究在简要分析逐级正交匹配稀疏分解(OMP)和稀疏分解匹配追踪算法(MPA)的基础上,利用设定带噪信号粒子群与子字典之间内积的固定阈值形式改进MPA算法收敛速度慢的缺陷。利用信号噪声比和均方误差对改进后的正交匹配追踪算法进行去噪和识别的检验,以此判断输电信号的识别效果。通过对改进后的算法进行实验仿真表明:在信号的去噪过程中,MPA算法最接近原始输电信号,而改进后的正交匹配追踪算法的波形幅度最小且曲线更为平滑,信号噪声比为27.18,均方误差为0.0209;在信号的后期识别过程中,仅通过20个以上采样数据点的位置即可有效的识别出带噪信号。
For transmission signal by ambient noise signal disturbance problems identified in the course of long-distance transmission, a brief analysis of the present study in stepwise orthogonal matching sparse decomposition （OMP） algorithm （MPA） and based on sparse decomposition matching pursuit, with the use of set fixed threshold value in the form of the inner product between the noise signal and the sub-dictionary particle swarm algorithm improved MPA slow convergence defects. And using the signal to noise ratio of the mean square error orthogonal matching pursuit algorithm improved denoising and identification tests in order to determine the effect of identifying the transmission signal. Through the improved algorithm simulation experiments show that： in the process of de-noising signal, MPA algorithm closest to the original transmission signal, and the waveform amplitude improved orthogonal matching pursuit algorithm is the smallest and smoother curve, the signal to noise ratio 27.18, mean square error of 0.0209; late in the identification process signals, only 20 more than the location of the sampling data points can effectively identify the noisy signal.
出处
《电子设计工程》
2017年第7期161-164,169,共5页
Electronic Design Engineering
基金
国家自然科学基金项目(60673153)
国网山东省电力公司科技项目(520612160004)
关键词
逐级匹配
稀疏分解
匹配追踪
输电信号
信号识别
stepwise match
sparse decomposition
matching pursuit
transmission signal
signal recognition
