The Study On Constant Modulus Algorithms Of Beam-forming

Blind adaptive beam-forming algorithm has become the research emphasis of many specialists and scholars, because of its merits requiring no reference signal. This paper focuses on researching one kind of important algorithm in the blind beam-forming—the constant modulus algorithm(CMA). "Blind" algorithm is developed on the basis of the traditional adaptive beam-forming algorithm, therefore we need to study CMA by comparison with the traditional adaptive algorithms to analyze the advantages and disadvantages. The constant modulus signal is one kind of the most important communication signals, especially in CDMA systems, where most signals are constant modulus. And the constant modulus algorithm which is used specially in the constant modulus signals, has so many marked merits such as easily computable, simply realizable, insensitivity to the error of the adaptive array model. So CMA has drawn more and more attention of the scholars of the world.From the three aspects of algorithm theory, mathematical model, algorithm programming and simulation, this paper studies systematically the convergence behavior of CMA emulation, showing that the performance of CMA is good. However, the performance of CMA is affected by SNR, and the convergence rate is decided by the step-size factor. When the step-size factor is badly chosen, it will lead to divergence consequence. Therefore, we proposed a variable step-size constant modulus algorithm by analyzing the effect of the iterative step-size in the iterative optimization process. Simulation results show that the variable step-size constant modulus algorithm has faster convergence rate. At the same time, this paper analyzes the advantages of constant modulus algorithm CMA(2, 2) and CMA(1, 2) in the convergence rate and proposed a modified constant modulus algorithm by combining the advantages of these two algorithms. The simulation results also verified this improved algorithm has faster convergence rate.At the same time, we should clearly know that CMA and other blind beam-forming algorithms are still in development. Though they could not be used in real adaptive array systems now due to the speed of system processing, it is no doubt that its prospects are good.