Research Of Adaptive Beamforming Algorithm In Smart Antenna

The channel of wireless mobile communication is a multi-access and multi-user channel, and it has some problems, such as, multi-path fading, time delay spreading, frequency spreading. At present, conventional technologies which are used to solve these problems include channel coding, equalization, diversity etc, although these solutions become mature, they can't solve these problems very well. Smart antenna and space filtering have brought new ideas to the solution of solving these problems, and smart antenna is considered as the last border of wireless communication technology. The core of smart antenna technology is the adaptive beamforming algorithm. Different adaptive beamforming criterions are used for different signal and context, the optimum results of these different criterions can be decomposed to the product of the same linear matrix filter and the different scalar processor, and all the optimum results converge to the optimum results of Wiener-Hope equations. It does not have meaning of the decision to select some adaptive beamforming criterion for using, However, choosing which kind of adaptive beamforming algorithms to adjust the directional pattern of smart antenna is very important, because these algorithms determine temporarily respond speed and the complexities of circuit of smart antenna.Firstly this paper has introduced the basic concept, structure and four kinds of adaptive beam forming criterion of the smart antenna, they are the maximal Signal to Noise criterion, the minimum mean square error criterion, the least square criterion and the linear constraint minimum variance criterion, and make comparisons between them on cost function, the optimum result, advantage and disadvantage, etc. Then this paper do further investigation four kinds of adaptive beam forming algorithms of smart antenna, they are the least mean square algorithm, the recursive least square algorithm, the least square constant module algorithm and the linear constraint minimum variance algorithm. LMS algorithm has simplicity of implementation and robust performance, but its convergence behavior mainly affected by the step-size parameter and the correlation matrix of tap-input vector. RLS algorithm has faster convergence speed and smaller excess mean error than LMS, but RLS has more complicated computation than LMS algorithm. Although LS-CMA algorithm has worse convergence performance then LMS or RLS, LS-CMA does not need reference signal and can be implemented by its constant module inherently. LCMV is constrained to pass the target signal with linear response, while at the same time minimizing the total output variance or power by adjusting the weight vector. This paper constructs the emulation platform of smart antenna by MATLAB, emulates the learning curve of performance about these different algorithms, discusses the advantage and disadvantage, applicable context of these them, and gives the contrast learning curve of performance between them.At last the paper simply introduces the two-stage signal processing structure of smart antenna which can easily solve the restriction of calculation improve the performance of system.