Research On Noise And Sparse Channel Estimation Method Of Low Voltage Power Line Communication

Power line communication?PLC?relies only on the power line network as the communication medium to realize broadband and high-speed communication and has become one of the most potential communication solutions in the emerging field such as smart grid and smart home.However,the power line which is originally designed for power transmission is harsh for high frequency signal propagation and has severe frequency selective fading and complex noise.In order to ensure the communication reliability of PLC,research on power line channel is an important orientation to overcome the inherent defects.The least square?LS?based conventional channel estimation method has not made full use of the power line channel characteristics,which makes it suffer from a low estimation accuracy,pilot overhead and low spectral efficiency,and hence can't achieve the optimal estimation results.Fortunately,the compressed sensing?CS?based sparse channel estimation method can overcome the drawbacks in the conventional method and has become an alternative which will make contributions both to academic research and engineering application in the power line channel estimation.The main conducted research work and conclusions in this dissertation are organized as follows:1.Noise is studied in the power line channel and simulations is conducted for each noise components.The results reveal that the impulsive noise has more severe interference on the PLC communication quality due to its strong randomness and high energy.Therefore,impulsive noise mitigation becomes the main design objective of communication anti-jamming.2.The power line channel characteristics is studied in the single-input single-output?SISO?case and multiple-input multiple-output?MIMO?case in terms of their multipath transfer function.And the simulation results reveal that the power line channel has frequency selective fading.3.A novel concept called“parametric sparsity”is introduced into the power line multipath channel model.The sparse channel estimation model which has the same mathematical model as the sparse signal reconstruction can be deduced by exploiting the sparsity of power line channel.Through this rigorous process,the power line estimation can be transformed into a CS-enabled sparse signal reconstruction both in SISO case and MIMO case.4.A relevance vector machine?RVM?based channel estimation algorithm is studied for SISO PLC channel estimation.And the simulation experiment results indicate that this algorithm has better mean square error?MSE?performance than other algorithms.Specifically,the MSE can reach to 4.5×10^-3 when the signal to noise ratio?SNR?is 20dB.Moreover,its spectral efficiency can reach to 87.4%which is 10%greater than the conventional LS based algorithms.The simulation experiment demonstrates the effectiveness of RVM algorithm applied into SISO PLC channel estimation.5.A block sparse Bayesian learning?BSBL?based channel estimation algorithm is studied for MIMO PLC channel estimation.And the simulation experiment results indicate that this algorithm reduces pilot overhead and has better MSE performance than other algorithms.Specifically,the MSE can reach to 1.8×10^-3 when the SNR is 20dB.The simulation experiment demonstrates the effectiveness of BSBL algorithm applied into MIMO PLC channel estimation.