Research On The Blind Beamforming Algorithms Based On Constant Modulus Property For CDMA System

Smart antenna has been widely researched in communication and signal processing. The adaptive beamforming algorithm is the key technique of smart antenna. It could adaptively steer the array directional pattern to form the main narrow beams with high gain in the direction of the users' signals while deeply nulling the interference signals. It is an effective way to realize the optimal receiving of the interesting signals. However, the additional adaptive beamforming algorithms need some kind of reference signal, and transferring the reference signal would occupy some frequency which is valuable and in short. Therefore, the blind beamforming algorithms that does not need reference signal have been developed as a research hotspot these years. This paper focuses on an important kind of blind beamforming algorithm - the constant modulus algorithm (CMA), and carries out a number of studies. Moreover, the expanding algorithm of CMA is applied to the CDMA system, and gets some meaningful results. The author's main contributions are as follows:1. The problem of smart antenna beamforming is analyzed seriously. Several adaptive beamforming algorithms are simulated, and the performances of these algorithms are compared, also we point out the advantages and disadvantages of these algorithms.2. The smart antenna signal model for Gauss channel is given. One kind of multi-target blind beanforming algorithms - the kind of multi-target despread-respread algorithms which are applicable to the CDMA system are discussed and analyzed. Furthermore, we present an improved least square despread-respread multi-target constant modulus algorithm (LS-DRMTCMA), which can adjust the coefficients of LS-DRMTCMA.3. From above work, a new blind beamforming algorithm - code filtering least square constant modulus algorithm (CF-LSCMA) is presented. The simulation results indicate that this algorithm have several advantages such as fast convergence, stable performance and so on, it can also complete multi-user separation accurately. Compared with the classic multi-target despread-respread algorithms, CF-LSCMA has better performance.