Research On Blind Equalization Algorithm For Multi-input Multi-output System

In MIMO communication system, there exist Co-channel interference, ISI, and other factors, which cause signal distortion, so TI is important to process the receive signal by using equalization technology. However, the blind equalization algorithm, which does not need to send training sequence, will save channel bandwidth and improve the signal transmission rate effectively. MIMO techniques have gained more and more interest in the last few years. As information theory indicates, large capacity is obtained via the potential de-correlation between the channels. So it is significant to improve the traditional blind equalization algorithms in single-input single-output (SISO) system and apply them into the MIMO channel. In this paper, the main research work is as follows:(1) Firstly, principle of Support Vector Machine is studied and how Support Vector Machine affects the speed of algorithms convergence is analyzed, cross-correlation combining with Energy Normalized Cross-correlation Constant Modulus Algorithm, a constant Modulus Algorithm Based on SVM Initialization is proposed. Then, getting the optimal parameter ranges with further researching on Forgetting Factor and Mixing Parameters effecting convergence rate of algorithms. Finally, with the researching of advantages of Orthogonal Wavelet in accelerating the convergence rate, Orthogonal Wavelet based Energy Normalized Cross-correlation Constant Modulus Algorithm is proposed, and then the performance of these two algorithms is compared.(2) Mixed square contour algorithm Applied in MIMO system is proposed by researching the influence on convergence rate by using different judgment domain to be the switching Criteria, incorporating which into the MIMO system and combining with the Energy Normalized Cross-correlation Constant Modulus Algorithm. Then, based on Variable Modulus Square Contour Algorithm, Variable Step-Size Variable Modulus Square Contour Algorithm is improved and applied in MIMO system; the algorithm can improve the convergence speed, and has the function of carrier phase recovery by using different step-sizes in different domains. Finally, combined with Autocorrelation Error Function, Variable Step-Size Variable Modulus Square Contour Algorithm Based on Multi-delayed Autocorrelation Error Function is proposed, this algorithm can effectively speed up the convergence and enables smooth and stable convergence curve because of its automatically adjusting of step-size with the error function.(3) Multi-mode Blind Equalization Algorithm can correct phase rotation, but which convergence is slow; an improved Weighted Multi-modulus Blind Equalization Algorithm is proposed and incorporated into the MIMO system. Then, Variable Step-Size Weighted Multi-modulus Blind Equalization Algorithm is proposed based on P-VAVSCA algorithm; this algorithm accelerates the convergence rate by changing the fixed step-size into variable step-size. Finally, a Multi-mode Blind Equalization Algorithm for fast QAM signals is researched and incorporated into the MIMO system; this algorithm can accelerate the convergence rate and has a small amount of computation..(4) Study the Matlab programming language, integrate the blind equalization algorithm in the original SISO systems proposed in this paper for MIMO systems, and write Matlab program to build a simulation platform, the simulation platform can provided simulation of different algorithms, you can quickly achieve a variety of signals simulation in different environments.