Research On Channel Estimation Of Low-voltage Power Line Communication System Based On OFDM

The application of low-voltage power lines for home broadband access has the advantages of convenient laying and low cost,so power line communication technology has received increasing attention.However,the low-voltage power line does not consider the transmission of high-speed digital signals,since it was originally designed to transmit power.This makes it possible to encounter various additional problems,such as multipath fading and severe noise interference,when using a low-voltage power line as a signal transmission medium.In order to solve the above problems,OFDM technology has become the mainstream modulation technology of low-voltage power line communication systems.For an OFDM system,the channel estimation module is critical and has a significant impact on the effectiveness and reliability of the entire low-voltage power line communication system.This paper has a research on the channel estimation of low-voltage power line communication system based on OFDM.Firstly,the basic characteristics of low-voltage power channel are introduced.The impedance characteristics,attenuation characteristics,time-varying characteristics and noise characteristics are analyzed in detail.Then the basic principles of OFDM technology is introduced.At the end of this section,the advantages and disadvantages of OFDM modulation technology applied in PLC field are analyzed.Secondly,the pilot-assisted channel estimation algorithm is studied,and the principles of several classical channel estimation algorithms are described.An improved LS channel estimation algorithm is proposed in this paper in order to solve the problem that LS channel estimation algorithm is not suitable for channels with impulse noise interference.This improved algorithm uses Pseudo-Huber function with better suppression effect on excessive residuals,which can effectively suppress interference of impulse noise.The simulation results show that the bit error rate and mean square error of the researched algorithm are both smaller than the result of LS channel estimation algorithm.Thirdly,the threshold-based DFT channel estimation algorithm is studied.The threshold setting will be too large when the pulse noise component appears outside the length of cyclic prefix.Because of the too large threshold setting,the effective component that during the length of cyclic prefix will be filtered out by mistake.Two improved algorithms are proposed in this paper based on the above problems.One is to determine the type of noise component based on the ratio of whole noise amplitude to the median of the overall noise amplitude,and then filter out the impulse noise component.In the other algorithm,the impluse noise components outside the cyclic prefix length are discriminated and rejected by Grubbs detection.The simulation results show that the channel estimation effects of the two improved algorithms are better than the threshold-based DFT algorithm.