Research On XPIC Technology Based On Blind Source Separation

Aiming at the problem of spectrum congestion in satellite communication,introducing polarization information to improve the system frequency utilization becomes an effective way to solve this problem.With the wide application of dual-polarized antennas and phased array antennas in satellite communications,polarization information processing technology that utilizes dual-polarization to achieve spectrum multiplexing has emerged as the times require.Using dual-polarization multiplexing technology can improve spectrum utilization and effectively increase system capacity.However,due to the influence of factors such as channel non-ideal characteristics,rain attenuation,and antenna isolation,the polarization state of the transmitted signal changes.The two signals transmitted in the cross-polarization mode leak each other,causing cross-polarization interference(XPI)and affecting the performance of the communication system.Therefore,in the satellite high-speed data transmission system using the dual-polarization working mode,it is necessary to deal with the problem of the decrease of signal-to-interference ratio caused by XPI.This paper studies the cross-polarization interference cancellation(XPIC)technology.Firstly,the paper introduces the polarization principle of electromagnetic waves,establishes a XPI transmission model,analyzes the existing cross-polarization interference cancellation technology,and lays a theoretical foundation for subsequent research.Since the traditional correlation cancellation XPIC algorithm fails to consider the phase shift effect caused by XPI,this paper takes this effect into consideration and proposes an XPIC optimization algorithm.The algorithm does not use additional training sequences,but calculates the observed signal as the expected signal.The applicable conditions and cancellation performance of the algorithm are analyzed through theoretical derivation and simulation.Secondly,since the cost function of the traditional blind adaptive CMA algorithm only includes the signal amplitude information and does not take into account the phase information of the signal,this paper uses the MCMA algorithm to eliminate XPI,and divides the signal into I and Q two channels for interference cancellation.Through the simulation comparison with the traditional CMA algorithm,the improved CMA algorithm has better convergence performance and higher signal to interference plus noise ratio(SINR),which provides some references for the blind adaptive algorithm in the field of XPIC.Finally,considering that the traditional XPIC method relies on training sequences or additional auxiliary information,resulting in waste of spectrum resources,this paper proposes a XPIC algorithm based on RobustICA,using the kurtosis comparison function Exact linear search optimization is performed to select the optimal step size,in this way any independent component with non-zero kurtosis can be selected,and the ordering problem in blindly separated signals is solved.The algorithm can separate the source signal from the observed signal without prior knowledge of the signal and subsequent demodulation,so as to eliminate the polarization interference.Compared with FastICA’s XPIC algorithm,it enhances the robustness when saddle points and pseudo local extreme points appear in the signal,and reduces the overall complexity.In view of the fact that the existing cancellation method after demodulation depends on the successful demodulation of two signals,in the case of strong XPIC,the cancellation may fail.This paper mainly studies the cancellation method before demodulation,optimizes the current XPIC algorithm,and proposes to apply the blind separation algorithm to XPIC.The feasibility and effectiveness of the algorithm proposed in this paper are verified by simulation.Compared with the current algorithm in the field of XPIC,the algorithm in this paper has better performance.