智能电网窄带OFDM通信系统噪声抑制

发布时间：2018-06-04 09:19

本文选题：电力线通信 + 智能电网　；参考：《电工技术学报》2014年11期

【摘要】：电力信道在低频处具有高的噪声,严重影响电力线通信(PLC)系统的性能。针对窄带(9~95k Hz)电力信道非平稳分布的脉冲噪声和平稳分布的背景噪声,根据实际测量电力噪声的统计分析,建立随机分布特性参数所描述的噪声模型,并提出相应的噪声抑制(Noise Suppression,NS)方案。其中脉冲噪声使用限幅(Clipping)方法进行消除,这是一种无记忆非线性技术;背景噪声因具有高斯特性,而正交频分复用(Orthogonal Frequency Division Multiplexing,OFDM)信号的正交和同相分量均符合线性分布模型,利用KLT(Karhunen-Loeve)变换将含噪信号协方差矩阵的特征值分解至噪声和信号子空间,在信号失真最小准则前提下,应用拉格朗日最优极值法,推导OFDM信号的线性估计。基于G3-PLC物理层和窄带噪声模型,构建OFDM仿真系统,对窄带OFDM系统噪声抑制前后性能的仿真,表明该噪声抑制方案,能够改进BER性能,提高SNR。
[Abstract]:The power channel has high noise at low frequency, which seriously affects the performance of power line communication (PLC) system. In view of the non-stationary impulse noise and the stationary background noise of the narrow-band power channel, the noise model described by the random distribution characteristic parameters is established according to the statistical analysis of the measured power noise. The corresponding noise suppressionNSs scheme is proposed. The pulse noise is eliminated by clipping method, which is a memoryless nonlinear technique, and the orthogonal and in-phase components of orthogonal Frequency Division Multiplexing (OFDM) signal are in accordance with the linear distribution model because the background noise has the characteristics of Gao Si. The eigenvalues of the noisy signal covariance matrix are decomposed into noise and signal subspaces by KLTT Karhunen-Loevetransform. The linear estimation of OFDM signals is derived by using the Lagrangian optimal extremum method on the premise of the minimum signal distortion criterion. Based on the physical layer of G3-PLC and the narrowband noise model, a OFDM simulation system is constructed. The simulation results show that the proposed scheme can improve the performance of BER and improve the performance of SNR.
【作者单位】：湖南大学电气与信息工程学院;湖南科技大学物电学院;
【基金】：国家自然科学基金(61172089) 高等学校博士学科专项科研基金(20120161120012) 湖南省科技计划项目——国际与区域科技合作(2014WK3001)资助项目
【分类号】：TM73

【参考文献】