基于OFDM的电力线载波通信信道估计

发布时间：2018-03-06 01:20

本文选题：电力线载波通信　切入点：OFDM　出处：《吉林大学》2017年硕士论文　论文类型：学位论文

【摘要】：电力线载波通信(Power Line Communication,PLC)是指利用已有的电力线网络进行信号传输的通信方式,具有使用用户多、传输距离远、通信可靠性高、不需要重新布线、电网建设同步等优点。电力线系统由于实际使用环境的复杂性,信道信息会受到多径衰落的影响,严重影响了系统性能效果的实现,通常情况下我们采用信道估计的方法来对信道的响应变化进行判断和分析,但与此同时,信道估计方法精确的程度会对系统性能产生影响,所以我们一直在不断追寻更好的解决方案。本论文主要研究的是基于OFDM的电力线载波通信信道估计方法。首先介绍了电力线载波通信现阶段的研究背景与意义。再简单描述了国内外PLC发展现状及信道估计的意义,然后介绍了电力线载波通信的基本原理,并分析了电力线信道特性,包括时变特性、衰减特性、阻抗特性和噪声特性。后对电力线载波信道进行建模,并用MATLAB模拟了4径和15径情况下的信道模型。介绍了基于OFDM的信道估计方法,并给出了经典LS算法、LMMSE算法和SVD算法的推理和仿真结果,为后文提出的方法作对比项。分析了高斯过程(GP)和高斯回归过程(GPR)的基本定义。给出了基于LS-GPR的电力线信道估计过程并通过MATLAB仿真得出结果。其中,算法中利用LS准则进行估计时并没有考虑到收到信号与原导频信号之间的关系及噪声的影响,因此,为提高信道估计的系统性能,采用了MMSE来降低噪声的影响。我们分析得知噪声子空间的白噪声满足均值为零的高斯分布,子载波间的干扰也满足均值为零的高斯分布,从信道信息可以分析得出有用的信道冲激响应变化相对于噪声分量的变换是十分缓慢的,在这里我们可以充分利用IFFT变换将有用信息和想要舍去的噪声分量分开。并用MATLAB实现理论分析,得到结论。综上,本文采用的最小二乘高斯回归过程的信道估计方法。高斯回归过程在贝叶斯准则下利用概率方法学习数据,具有非常强大的非线性拟合能力和小样本学习能力,并可以由一个最优化过程自动获取最佳超参数。本文提出两种基于LS-GPR信道估计算法的改进方法,运行结果证明了方法的有效性。
[Abstract]:Power Line Communication (PLC) is a communication mode which uses the existing power line network for signal transmission. It has many users, long transmission distance, high communication reliability and no need for rewiring. Because of the complexity of practical application environment, the channel information will be affected by multipath fading, which seriously affects the realization of system performance. In general, we use channel estimation method to judge and analyze the channel response change, but at the same time, the accuracy of channel estimation method will affect the performance of the system. Therefore, we have been looking for a better solution. In this paper, we mainly study the power line carrier channel estimation method based on OFDM. Firstly, we introduce the background and significance of the power line carrier communication research at the present stage. It also briefly describes the current situation of PLC development and the significance of channel estimation at home and abroad. Then the basic principle of power line carrier communication is introduced, and the characteristics of power line channel, including time-varying characteristic, attenuation characteristic, impedance characteristic and noise characteristic, are analyzed. The channel models with 4 and 15 paths are simulated by MATLAB. The channel estimation method based on OFDM is introduced, and the reasoning and simulation results of classical LS algorithm and SVD algorithm are given. The basic definitions of Gao Si process and Gao Si regression process are analyzed. The power line channel estimation process based on LS-GPR is given and the results are obtained by MATLAB simulation. In order to improve the system performance of channel estimation, the relationship between received signal and original pilot signal and the influence of noise are not taken into account when LS criterion is used in the estimation. MMSE is used to reduce the noise effect. We find that the white noise in the noise subspace satisfies the Gao Si distribution with zero mean value, and the inter-carrier interference satisfies the Gao Si distribution with zero mean value. From the analysis of channel information, it can be concluded that the change of channel impulse response relative to noise component is very slow. Here we can make full use of the IFFT transform to separate the useful information from the noise components that we want to lose. We can use MATLAB to realize the theoretical analysis and get the conclusion. In this paper, the channel estimation method of the least squares Gao Si regression process is used. Under the Bayesian criterion, Gao Si regression process uses the probability method to learn the data, which has very strong nonlinear fitting ability and small sample learning ability. In this paper, two improved methods based on LS-GPR channel estimation algorithm are proposed, and the results show that the method is effective.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN913.6

【相似文献】