Study On Smart Antenna Algorithms For Wireless Communications

The smart antenna has been the one of key techniques in future wireless communications, which combines the antenna array technology with the signal processing in space and time . The spatial processing leads to more degrees of freedom in the system, improves its overall performance, and enhance the system capacity and spectrum efficiency. In this dissertation, a comprehensive study on adaptive algorithms of smart antennas in the CDMA and OFDM wireless systems, especially, the adaptive algorithms using the neural network is presented with simulated results, which mainly includes four parts stated as follows.Firstly, Some different non-blind adaptive beamforming algorithms of smart antennas are simulated , then their performances and features are compared. For the CDMA system, a improved algorithm with relatively lower computation complexity and good restraining interfere performance is proposed. Moreover, a adaptive power control algorithm for the CDMA system is simulated and compared with a conventional method. For the OFDM system, a scheme of combining space-time code with the adaptive power control technology is proposed with simulation results, which increases the diversity gain and the adaptive gain.Secondly, a scheme using MMUSIC (Modified Multiple Signal Classification) algorithm for the wideband coherent source DOA(Direction of Arrival) estimation is introduced, in which the calculation method of applying average means is adopted to enhance the space spectral estimation precision. Especially, a DO A estimation approach by the radial basis function neural network is suggested, and the simulation results are shown, confirming the validity of this method.Thirdly, the separation of signal by an adaptive neural network method are described, which is suitable for the nonlinear noise dispose, resulting in good agreement between the simulated signal and the original source one. Also, an adaptive neural network blind...