The Study Of Adaptive Beamforming Of Smart Antenna

As a very promising research for wireless communication technologies, smart antenna is now used in communication systems successfully. It uses modern digital signal processing techniques and selects appropriate adaptive algorithms to form space directional beam dynamicly, which taking full advantage of mobile users to achieve the purpose of offsetting or suppressing interference signals to the greatest extend. This thesis is mainly focused on the core content of smart antenna, which is the adaptive beamforming algorithms, including:First of all some kinds of beamforming algorithms are introduced, including non-blind algorithms, algorithms based on DOA and blind algorithms, and their performance is analyzed and compared with each other. On the basis of research, an improved algorithm based on DOA is proposed, which can reject interference strongly in view of projecting the previous weight vector to the structed orthogonal sub-space of the interference signals. Then based on the problems of the constant modulus algorithms capturing the interference signals and the preprocessing algorithm applied in corruptive environment, a widely used constant modulus algorithm based on minimal Rayleigh entropy is proposed. The simulation results indicate their effectiveness.Secondly, the beamforming techniques for CDMA systems are introduced emphatically. Through the traditional adaptive beamforming is achieved before spreading signals, which is limited in the environment of high chip rate. In addition, the reference signal needs respreading, the system is further complicated. Hence the approach based on code filtering is applied, with the local spreading codes multiplied the received signals to despread for beamforming. And based on the maximum signal to interference and noise ratio, an integrated method that using an on-line iterative algorithm with low complexity in the code filtering beamforming is proposed. Simulation results confirm the effectiveness of the algorithm that can be applied in different communications environment.Finally, the thesis discusses several different kinds of blind multi-user separation algorithms, which include multistage separation, Gram-Schmidt orthogonalization and cost function transformation algorithms. In view of these algorithms existing the shortcomings of complexity and instability convergence, a new algorithm based on the iteration least squares projection constant modulus algorithm is proposed in this chapter. It can separate the independent signals from different directions effectively without knowing the transcendent knowledge of the signals. And the proposed algorithm don't need to use Gram-Schmidt orthogonalization procedure, and can estimate the users' DOA on the basis of separation. The computer simulation confirmed the algorithm not only is simple and can converge quickly, but also has the advantages of robustness and effective separation.