MMSE based FDD overhead optimization for a multiuser two-way system

Based on maximizing the sum of lower bounds derived for the uplink (UL) and the downlink (DL) rates with imperfect Channel State Information (CSI), overhead optimization i.e. finding the optimum lengths of the UL and DL pilots and feedback bits, is performed. Estimated CSI and quantized Channel Direction Information (CDI), which may be received erroneously, are available at the Base Station (BS) for computing the receive and the transmit Wiener filters (WF), respectively. Using large system analysis, we were able to derive analytical DL and UL mean squared error (MSE) expressions that give a lower bound of the DL and UL rates, respectively, needed for our optimization.