Research On Blind Equalization Algorithm Based On Signal Correlation For MIMO System

In MIMO communication system, research on all kinds of blind equalization techniques has become a key issue, blind equalization algorithms could compensate signal distortion caused by the various types of interference.Because of the unique co-channel interference and inter-user interference of the MIMO system,there is some correlation between the received signals.Using the traditional MIMO signal equalization algorithms in signal processing, the steady-state error is large, a plurality of the equalizer output signals may also be locked to the same emission signal. This paper introduces the basic features and blind equalization of MIMO system on the basis of theoretical knowledge, blind equalization algorithm of MIMO signal is studied,the main work is as follows:(1) Study the characteristics of the cascading multiple-input single-output equalizers,and used it for the bind equalization for MIMO signals.Take the cascading MISO-CMA as the example, simulation results demonstrated the effectiveness of the MISO equalizers in processing the signal correlation. Then use the cascading MISO equalizers structure into the combination algorithms as CMA+DFE and CMA+AMA,experiments show that the combination algorithms had greatly decreased steady-state error due to the reducing influence of correlation between MIMO signals in the equilibrium process.On this basis, the dynamic changes of weight factor is introduced into the CMA+AMA combination algorithm to control the CMA and AMA algorithm right combination of weight change in the algorithm to achieve a better convergence results.(2) Firstly, a multi-modulus blind equalization algorithm based on linear cross-correlation item was proposed, the study found that the algorithm could de-correlate the correlation between the mixed MIMO system receiving signals,but the solution ability is weak. This paper presented a new nonlinear cross correlation term, which owning a stronger de-correlation capability using the autocorrelation function of the output signals of MIMO equalizer to do a nonlinear processing,so the multi-modulus blind equalization algorithm based on the nonlinear cross-correlation item performed better than the linear one. In order to further accelerate the convergence speed without affecting the steady-state error, the conjugate gradient multimode algorithm with dynamic momentum factor changing based on nonlinear cross-correlation item was improved, and the effectiveness of the improved algorithm was proved by simulation experiments. (3) The weighted multi-modulus blind equalization algorithm based on principal component analysis(PCA) for MIMO system was proposed. Principal component analysis could de-correlate the correlation and reduce the dimensionality. Using PCA for signal preprocessing for MIMO system, both to decrease the convergence error because of reducing the correlation between the receiving signals, and improve the convergence speed by lowering the subsequent peacekeeping operations blind equalization algorithm. On the basis of researching on the characteristics of the traditional PCA algorithm, the weighted multi-modulus blind equalization algorithm based on the under-determined PCA algorithm and the MDL-PCA algorithm were proposed,in addition,using them for the pretreatment of MIMO signals, then simulation experiments through MIMO channel model proved their feasibility and validity.