Study Of The Blind Equalization Algorithm For High-order QAM

QAM(Quadrature Amplitude Modulation)is a kind of high spectral efficiency modulation. But in the actual transmission process, due to the influence of the time variant of channel, the ideal transmission characteristics can not be obtained. And it will produce waveform distortion, inter symbol interference(ISI) and inter Channel interference(ICI). While the equalization techniques can correct and compensate the system characteristics, and reduce the ISI and ICI in a certain degree.Conventional equalization techniques update the equalizer tap coefficients iteratively through a training sequence in the channel known the characteristics. After the finite iterations, the equalizer may approach the ideal state. However, sending training sequence requires additional channel bandwidth which make the channel resource not enough. So it is important to study the blind equalization without knowing the channel information and the training sequence under the situation that we know a priori information.In this paper, we study the characteristics of high order QAM signals, and the corresponding blind equalization algorithm. The work done by the following points:Firstly, the core operating mode of this kind of blind equalization algorithm is introduced by the description of some properties of the Bussgang blind equalization algorithm. And several important Bussgang blind equalization algorithms are introduced.Secondly, the most common and basic constant modulus(Constant Modulus Algorithm, CMA) algorithm is introduced. Through the analysis of its performance, it is pointed out that the CMA algorithm has the limitation of correcting the phase rotation and causes the large error due to the fixed step size.Thirdly, comparing the multi mode blind equalization algorithm with the traditional CMA blind equalization algorithm according to the characteristics of high order QAM signals. Then the CMA algorithm and the multi model algorithm are fused, which can improve the performance of the algorithm in a certain extent. But the algorithm still can not solve the two problems above. And also in the project I participated in this signal needs to be equalized.In the case that the signal to noise ratio is 15 dB, the modulation method of the signal should be recognized with the expected rate can reach 90%. The performance of the equalization has a direct influence on the correct rate of the modulation recognition. However, the output SNR of the algorithm is about 30 d B, which can't meet the above requirements, so this also forced me to further study the performance of the equalizer.Finally, after the analysis of the limitations of the algorithm and the requirements of the performance of the actual experimental platform, the variable step size of the dual mode multi- mode multi- mode blind equalization algorithm is proposed on the basis of the above algorithm. And the simulation results show that the new proposed algorithm can improve the performance of correcting the phase rotation and the fixed step size. And also through the experimental platform we can verify that the improved algorithm improve the signal to noise ratio of 50% or so, which meet the requirements of the experimental platform for data.