Research On Signal Detection Algorithm In MIMO-PLC System

Multiple-input multiple-output?MIMO?technology has been applied to power line communication?PLC?to meet the requirement of PLC for larger capacity and wider coverage to achieve high-speed data transmission.However,the MIMO-PLC channel has the characteristics of frequency selective fading,it causes the quality of different frequencies varies greatly,which makes the existing signal detection algorithms cannot achieve good performance.In addition,the performance of MIMO-PLC system is also affected by impulsive noise.As impulsive noise appears,the communication quality has been seriously degraded.In view of the above problems,the specific solutions are proposed in this thesis,and the contents are as follows:1.Considering that the large difference in channel quality between different frequencies,a detection algorithm based on condition number threshold selection is proposed in this thesis,the algorithm uses the number conditions of the channel matrix to measure the quality of the channel.The MMSE-SQRD?Minimum Mean Square-Error Sorted QR Decomposition?based on CLLL?Complex Lenstra-Lenstra-Lovasz?lattice basis reduction detection algorithm is selected when the channel quality is good,and the QRD-M?QR Decomposition with M-algorithm?detection algorithm is selected when the channel quality is poor.The simulation results show that the proposed algorithm can close the performance of optimal detection algorithm.In the 16QAM modulation mode,the complexity of the proposed algorithm is reduced by 44%compared with the QRD-M detection algorithm,and with the increase of modulation order,the complexity is reduced more obviously.2.The iterative impulse noise cancellation scheme with time-frequency combination is proposed for impulse noise seriously affects the quality of MIMO-PLC.The scheme is mainly used in transmit diversity mode,it uses the space frequency block code?SFBC?to improve the accuracy of hard-decision,and then after several iterations can minimize the influence of impulse noise through the simulation,the proposed algorithm compared with the traditional blanking algorithm,can improve its performance by about 2.7dB when bit error rate is10^-4 under the 16QAM modulation mode,and less iteration times,easy to be realized in engineering,and less iteration times,easy to be realized in engineering.3.Because the traditional signal detection algorithm can not be applied to the impulse noise environment of MIMO-PLC system,this thesis presents an improved maximum likelihood?ML?detection algorithm under impulse noise,and analyzes the characteristics of impulse noise distribution.The expression of ML detection algorithm is derived under this distribution.In order to reduce the complexity,two suboptimal detection algorithms are proposed in this thesis,which reduce the cost of exponential and logarithmic operations in the improved ML detection algorithm by piecewise function.The simulation results show that the performance of the three algorithms are significantly better than that of the traditional signal detection algorithm,and the performance of the three algorithms tend to be consistent.But considering the complexity,the two suboptimal algorithms are more suitable for the actual needs.