Optimal Linear Pre-coder and Power Allocation for MIMO-OFDM Systems with Partial and Imperfect Channel Knowledge

Multiple input multiple output (MIMO) systems provide dramatic capacity gain through an increased spatial dimension. However, the capacity gain is reduced if the channel state information (CSI) is not perfectly known. Thus, to improve MIMO system performance, CSI at the transmitter (CSIT) and receiver (CSIR) should be used efficiently. Uncertainties in CSI arise mainly due to imperfect channel estimation at the receiver and limited feedback capability of the reverse link. These two factors degrade the performance of practical MIMO systems. In this project we take into account these two effects and design optimal linear pre-coding and power allocation strategies accordingly. Several works in the literature have dealt with similar problems. In [1]-[3], the authors assume a perfect CSIR and limited feedback, and use transmit antenna correlation as partial channel knowledge to improve capacity. However, these works are restricted to the flat fading channel case. [4]-[6] study channel estimation error at the receiver and a transmitter power control to increase capacity. However, these results apply only to single input single output (SISO) channels. Other information theoretic treatments on MIMO capacity with imperfect CSI can be found in [7]-[11]. To consider the pre-coder design in MIMO frequency selective channel environments, we restrict our interests to the pre-coder design in OFDM systems because OFDM is the prevalent modulation scheme of wireless communication systems in frequency selective channel environments. General descriptions for OFDM systems can be found in [12]. Specifically, we consider two cases as follows.