Transmit Beamforming Weights for Channel Estimation Error and Feedback Latency in OFDM

This paper presents a new method for computing transmit beamforming weights in OFDM when there is channel state information (CSI) uncertainty at the base. The channel uncertainty may be caused by channel estimation error, quantization effects, and/or channel variations from when the CSI is measured and when the beamforming weights are applied. The weight calculation automatically trades off using long-term statistics (e.g., the average spatial covariance matrix) and short-term beamforming (i.e., maximal ratio transmission). Simulation results show that the new weights have very promising performance with the benefit of not requiring a means to decide whether the short-term or long-term beamforming should be used.