Interference Mitigation Methods For Low Voltage Power Line Communication Systems

Smart Grid systems perform fast,real time information exchange to monitor and control energy flows in order to improve the efficiency and reliability of power delivery.This monitor and control requires low-cost,low-latency,and highly reliable two-way communications infrastructure between customers and utility.Power line communications(PLC)have become an attractive communication solution for smart grid due to their ubiquity.Power lines,however,have traditionally been designed for one-directional power delivery and remain hostile environments for communication signal propagation.The power line network has complex topology and various access devices.The multipath effect makes power line become a frequency selective fading channel.In addition,there are a lot of bursty,non-Gaussian impulsive noises with high amplitude in power line channel.The hostile channel and rich impulsive noise are two main factors limiting the performance of power line communication systems.Orthogonal Frequency Division Multiplexing(OFDM)has been adopted in many modern communication standards as it can efficiently combat the intersymbol interference caused by the frequency selective channel.The presence of impulsive noise makes the performance of OFDM receivers degrade sharply.Aiming at the problem of the performance of power line communication systems restricted by the impulsive noise,the dissertation studies the impulsive noise mitigation and channel estimation in the presence of impulsive noise in order to design robust receivers.The main contributions of the dissertation are as follows:1?The shortcomings of two kinds of impulsive noise suppression methods are analyzed.One is the nonlinear preprocessor,which requires prior statistical information of impulsive noise.The other is to reconstruct the impulsive noise samples using compressed sensing technology based on the received signal in the null subcarrier of OFDM symbol.The performance of this method is limited by the number of null subcarriers.In order to reconstruct impulsive noise by using received signals in all subcarriers,based on the convex optimization estimation method of minimizing the impulsive noise vector in ?1 norm,we add data symbols as variables,and use received signals in data subcarriers as constraints.We construct a compressed sensing equation that can be used to reconstruct impulsive noise by using received signals in all subcarriers.In order to solve the non-convex optimization problem,the constellation point characteristics of data symbols using coherent modulation technology are analyzed.On the basis of known channel state information,the discrete constellation points are relaxed as continuous rectangular regions by relaxation principle,thus the non-convex optimization problem is transformed into a linear programming problem.In order to reduce the complexity of the existing linear programming algorithm,the alternating direction multiplier method is proposed to reduce significantly the computational complexity of the impulsive noise estimation method based on convex optimization.The simulation results show that the performance of the proposed impulsive estimation method using all subcarriers is better than that of the impulsive noise estimation methods using only the null subcarriers.2?Aiming at the deficiency of existing impulsive noise suppression methods which assume that channel estimation is independent of impulsive noise estimation,a joint estimation method of channel and impulsive noise based on sparse Bayesian learning theory is proposed.Based on the sparse characteristics of channel impulse response and time-domain impulsive noise,a sparse vector is constructed by connecting channel impulse response vector and impulsive noise vector.A compressed sensing equation for joint estimation of channel and impulsive noise is constructed by using received signals in null and pilot subcarriers,and a specific algorithm is implemented by using sparse Bayesian learning framework.In order to increase the performance of channel and impulsive noise estimation by using the received signals in data subcarriers,the transmitted data symbols are regarded as unknown hyperparameters.The iterative estimation of data symbols is realized by using the expectation maximization algorithm in the sparse Bayesian learning framework,and a joint estimation algorithm of channel and impulsive noise that can synchronously detect symbols is realized.In order to evaluate the performance of the algorithm,the Cramér-Rao Bound and system capacity expression of the proposed method are obtained.The simulation results show that the proposed joint channel and impulsive noise estimation method can significantly improve the performance of channel estimation and impulsive noise estimation.3?Based on the channel amplitude correlation of adjacent OFDM symbols,an estimation method which can exploit several OFDM symbols to estimate channel and impulsive noise recursively is proposed.Firstly,the channel impulse response tap coefficients of slow-fading channel or block-fading channel are modeled by using the first-order autoregressive model.The impulsive noise on each OFDM symbol is regarded as independence.Thus,the first-order autoregressive model is used to unify the channel and impulsive noise variation characteristics between adjacent OFDM symbols.By using Markov property,the joint estimation of channel and impulsive noise for multiple OFDM symbols is considered as a state estimation problem for linear dynamic systems.Because the dimension of the observed signal is much smaller than the dimension of the joint vector of channel and impulsive noise,the channel and impulsive noise estimation algorithm combining multiple OFDM symbols is realized based on Kalman filter and smoothing using the dynamic compressed sensing theory and the sparse Bayesian learning framework.The simulation results show that the method by using jointly multiple OFDM symbols can improve channel estimation performance and bit error rate performance more effectively than that of single OFDM,especially in block-fading channels.4?We develop a method to estimate the channel and impulsive noise that the channel impulse response is not sparse.The proposed channel and impulsive noise estimation method based on parameter sparsity by analyzing the classical power line multipath channel model.Firstly,the length of continuous channel transmission path is quantized,then the continuous frequency domain response vector of power line channel is sampled,and the frequency domain response vector is expressed as a linear transformation of sparse path parameter vector.Based on the sparse transform model,a method for estimating impulsive noise and power line channel for non-sparse channel impulse response is implemented.The simulation results show that this method is suitable for the actual power line channel model.