FDD-CDMA System Downlink Beamforming Researching

There is a phenomenon that the capacity of downlink is greater than that of uplink in the state of the mobile communication. It's seemingly prior to increase the capacity of uplink, but in 3G high speed data and multi medium business more occur in the downlink, so considering this kink of asymmetric requirement we can conclude that the bottleneck is downlink in future. It's very necessary and urgent to research the smart antennas in downlink.In forthcoming 3G mobile communication system, FDD system (like WCDMA, cdma2000) will be mainstream technology. Being different from TDD system, the behavior of up and downlink in FDD system is not so identical because of the difference of carrier frequency between uplink and downlink. So for the downlink of FDD system, beamforming is such a process that estimating the downlink channel covariance matrix from uplink first andthen allocating a weight vector for every user by a kind of optimization.This paper illustrates the principles of smart antenna and beamforming firstly, and thenanalyses the two difficult issues of smart antenna downlink beamforming when it is adopted in communicate with FDD, which are downlink channel covariance matrix estimation method and downlink beamforming algorithm. Secondly, the author have been studied these question associating with mobile channel characteristic and model. For downlink channel covariance matrix estimation, this paper proposes a novel method named frequency domain transformation, which is different from probing-feedback approach and DOA estimating method. The computer simulations show the proposed method has a better behavior of estimation and also reduce the complexity of computation. After getting the estimation of downlink channel this paper combines maximum SNR, general eigenvalue and power control conception to proposes a minimum comparative interference method in multi-cellular, which makes interference sum to in-cellular and adjacent cellular minimum when transmitting information for a objective user. The computer simulation result implicates these method can efficiently restrain interrupt and improve the system capacity.