Downlink Beamforming with Transmit-Side Channel Correlation: A Large System Analysis

In this paper, we consider a large system analysis of regularized channel inversion (RCI) beamforming for MISO broadcast channels (BC) with transmit-side channel correlation. In the analysis, we assume that both the number of users and transmit antennas grow unbounded with a constant ratio. We also assume that the channel correlation model is separable. Under this channel condition, we are particularly interested to find the optimal regularization parameter of RCI that maximizes the signal to interference plus noise ratio (SINR). First, we derive the large system limit of the SINR of each user by applying some results on large random matrices. Then, we determine the cor- responding optimal regularization parameter and show that this optimal regularization parameter is not affected by the transmit correlation. We verify this result through simulations where the channel has the exponential transmit-correlation model.