Blind Equalization Research Based On Non-cooperative Communication

In the process of signal transmission,the existence of Inter-Symbol Interference(ISI)will lead to the degradation of transmission quality,the receiver must recover the original signal as much as possible through corresponding processing,blind equalization technology plays an extremely important role in this process.Traditional adaptive equalization must provide training sequence,which is inefficient in transmission.Blind equalization overcomes this shortcoming,and the original transmission sequence can be recovered directly from the received signal and the equalized output signal.In this paper,various common blind equalization methods are thoroughly analyzed.The main research contents include:1.Classical Low-order Blind Equalization AlgorithmAlthough the original blind equalization algorithm implicitly uses fourth-order cumulants,it has been simplified,mainly using second-order statistics,the principle of the algorithm is simple,the calculation process is not cumbersome,and the application scope is wide.This paper mainly studies the commonly used Bussgang-type blind equalization algorithms(constant modulus,Goard and Stop algorithms)and their correction algorithms,including guided decision-least mean square(DD-LMS),minimum mean square error(MMSE),recursive least squares(RLS)and fractional interval(FSE)blind equalization algorithms.The characteristics of various algorithms are analyzed from the principle and implementation process gradually,and their performances are discussed based on the simulation results.2.Blind Equalization Algorithm Based on Higher-order Statistics AndHigher-order SpectrumIn order to make full use of the information in the signal,the higher order statistics and the higher order spectrum are introduced into the blind equalization method.Especially for non-Gaussian signal processing,it performs better,because it can effectively filter the Gaussian noise in the channel.This paper mainly studies the basic characteristics of higher-order cumulants and higher-order spectra,and on this basis,analyses the blind equalization algorithm of fourth-second-order cumulants and cepstrum initialization.Hyperexponential algorithm is a typical method usinghigher-order cumulants.Its convergence performance is improved at the cost of computational complexity.3.Blind Equalization Algorithm Based On Wavelet TransformWavelet transform uses dynamic window function to analyze the signals of each node,especially to extract local features of signals.The characteristics of window function correspond to time and frequency,and can form orthogonal basis.The characteristics of wavelet transform and its application in blind equalization are also studied in this paper.Based on the blind equalization method of single-channel wavelet transform,aiming at the shortcomings of time-consuming and error performance,a two-channel model is proposed and its implementation process is given.Compared with single channel model,the simulation results show that the proposed method performs better in performance.