MIMO OFDM capacity maximizing beamforming for large Doppler scenarios

Performance of OFDM systems is limited by inter carrier interference (ICI) under high Doppler scenarios such as that encountered in high speed trains such as TGV. Several publications have addressed the receiver design for SISO (single input single output) and SIMO (single input multiple output) to combat ICI. Notably, the use of multiple receive antennas is a very effective way to combat ICI. In this paper, we consider a MIMO (multiple input multiple output) scenario and iteratively design a transmit beamformer to maximize the sum capacity across all the subcarriers in the presence of ICI. Our interest lies in analyzing the impact of excess of transmit (Tx) antennas in combating ICI. A time varying frequency selective channel is considered. In particular, we consider a linear model for the channel variation across the OFDM symbol as we focus on LTE (long term evolution)-like systems and train velocities up to 450kmph. As the original cost function is non-convex, we first follow the difference of convex functions (DC) approach to obtain a convex cost function. However, we later reinterpret the DC approach as a majorization technique. To solve the joint problem of power allocation across subcarriers and precoder design for all the useful subcarriers, we employ the cyclic minimization approach to alternately optimize the precoder design and transmit power allocation across subcarriers. The convergence of the iterative approach is proved and the theory is validated via numerical simulations.