小波包结合奇异值分解在通信信道降噪中的研究  

Study on the Application of Wavelet Packet + Singular Value Decomposition in Noise Reduction for Communication Channel

作　　者：曹敏[1,2] 林聪[1,2] 李波[1,2] 刘清蝉[1,2] 梅二召 王行[3] 谢涛[3] 李川[3] 

机构地区：[1]云南电网有限责任公司电力科学研究院,云南昆明650217 [2]中国南方电网公司电能计量重点实验室,云南昆明650217 [3]昆明理工大学信息工程与自动化学院,云南昆明650500

出　　处：《自动化仪表》2017年第2期53-56,共4页Process Automation Instrumentation

基　　金：国家自然科学基金地区科学基金项目(51567013)

摘　　要：针对营配一体化通信信号受噪声干扰的问题,为了提高营配一体化通信信号的降噪效果,提出一种小波包结合奇异值分解的降噪方法。对电力线高速宽带载波信号进行小波包分解,有选择地对小波包分解系数进行奇异值分解和变换,并与小波包的多尺度时频分析及重构算法相结合,将其应用于通信信道降噪的研究中。针对不同的背景环境,挑选三个采集地点,采用TPS-2024数字信号示波器在现场采集数据,并对采集到的通信信号分别进行小波包降噪和小波包结合奇异值降噪。测试结果显示,小波包结合奇异值降噪较小波降噪的信噪比(SNR)提高10 d B左右,且均方误差(MSE)都有所降低。通过对比测试数据结果可知,采用小波包结合奇异值分解的降噪方法可以有效地保留信号的有效成分,降噪效果明显优于小波包降噪。To solve the problem of noise interference in communication of integrated marketing and distribution,and in order to improve the denoising effect of communication signals in integration of marketing and distribution,the method which is combining wavelet packet and singular value decomposition is proposed to reduce the noise. The high speed broadband carrier signals of the power line is decomposed by wavelet packet,and the singular value decomposition and transformation are selectively conducted for wavelet packet decomposition coefficients;and combined with the wavelet packet multi - scale time - frequency analysis and reconstruction algorithm, then applied in the research on noise reduction of communication channel. According to the different background environment,three of the collection sites are selected,using TPS-2024 digital signal oscilloscope to collect data in the field,and the communication signals collected are analyzed by denoising of wavelet packet and wavelet packet combining with singular value. The results show that the SNR of wavelet packet ＋ singular value de-noising can be improved about 10 dB,and the mean square error （ MSE ） is reduced. The results of comparing the test show that using wavelet packet ＋ singular value decomposition method can effectively keep the effective components of the signal,the noise reduction result is better than wavelet packet denoising.

关 键 词：电力 智能电网 通信 网络 载波 降噪 数据 

分 类 号：TH6[机械工程—机械制造及自动化] TP273[自动化与计算机技术—检测技术与自动化装置]