Antenna diversity techniques for orthogonal frequency division multiplexing (OFDM) wireless communications
This thesis consists of three topics related to MIMO (multi-input multi-output) OFDM systems. ^ First, pilot preamble schemes for OFDM systems with transmit diversity are proposed, and simple channel estimation techniques based on least-squares estimation (LSE) are provided. The first scheme is based on the locally frequency flat fading assumption. The pilot sequence for each transmitter antenna is the circular time shift of a pilot sequence, and the received signals at nearby subcarriers are used in the channel estimation. The local orthogonality in the received signal vector makes the estimation simple in terms of computational complexity. In the second scheme, pilot tones are placed sparsely in the frequency-space grid in a non-overlapping way, and the frequency correlation of the channel is used to estimate the channel. The complexity is reduced because fewer symbols per transmitting antenna are involved in the estimation. We analyze the estimation error and show the diversity gain and significant performance improvement of systems with transmit diversity using the proposed channel estimation schemes. ^ Second, we propose a simple receiver diversity scheme for OFDM systems. Diversity combining is done at the receiver before the discrete Fourier transformation (DFT) using a cyclic prefix. The scheme is shown to be equivalent to the conventional post-DFT maximal ratio diversity combiner. The computational complexity, or power consumption, can be reduced without any performance degradation when the number of antennas and the number of subcarriers are chosen properly. ^ Third, we propose OFDMA-CDM (orthogonal frequency division multiple access—code division multiplexing) systems with antenna polarization diversity, and analyze the impact of the channel correlation parameters on its average error-rate performance for various fading channels. We present closed-form average BER (bit error rate) expressions for frequency flat fading channels and a tight lower bound on the BER which is asymptotically achievable when there is independent and identically distributed (i.i.d.) fading at the subchannel level. The results are equally applicable to conventional transmit antenna diversity systems in correlated MIMO channels. The effect of nonorthogonal spreading sequences is also analyzed. ^
Major Professor: James S. Lehnert, Purdue University.
Engineering, Electronics and Electrical