Single user and multi-user transmission for OFDM with multiple antennas

The wireless channel has severe impairments, particularly Doppler and delay spread, which are due to time variation and multipath propagation. In order to combat Inter-Symbol Interference (ISI) caused by delay spread, Orthogonal Frequency Division Multiplexing (OFDM) is often employed. In addition, through the use of spatial diversity with multiple antennas, a wireless transmission system can enhance its performance dramatically. This spatial diversity can either increase the system capacity and/or cancel co-channel interference (CCI).; This dissertation studies the OFDM system with antenna diversity as a potential solution for high-data-rate wireless communications. In particular, the presence and absence of channel knowledge at the transmitter are investigated separately.; If the transmitter does not have channel information, the transmission can be optimized through the use of codes. With the space domain, a code can be designed to utilize spatial diversity as well as temporal diversity. We investigate coding strategies for an OFDM system with multiple antennas and compare their performances. For the receiver structure, we proposes a parametric channel estimation and MLSE detection algorithm.; If the transmitter has complete channel information, optimum data transmission is achieved through the use of a bit/energy loading algorithm. With multiple antennas, loading algorithm can be extended to two-dimensional processing to find the system capacity. This dissertation studies the capacity of a multiple antenna system and the impact of the number of paths in an outdoor wireless environment. With a space-frequency filter designed using Singular-Value-Decomposition (SVD), the transmission can be coordinated to obtain significant antenna gain.; We also describe a spatial multi-user access scheme using SVD (MU-SVD). MU-SVD uses spatial joint processing for CCI cancellation without losing antenna diversity. The proposed MU-SVD method outperforms conventional beamforming and space-time equalization. In addition, the MU-SVD method is less complex for adaptive updating when the channel is fading, more robust to the channel mismatch errors, and more near-far resistant than conventional CCI cancellation techniques. When signal signatures of two users are too close, MU-SVD can be modified to mitigate the singularity problem. For channels with delay spread, MU-SVD can also be applied to each subchannel in an OFDM system. In this scenario, we can appropriately allocate the power by a mixture method, water-filling for shared subchannels, or an on/off assignment for subchannels exclusively occupied by each user, according to the spatial signature correlation. In simulations, MU-SVD with power control achieves a suboptimum performance close to single-user bound.