| The use of multiple transmit and receive antennas can lead to significant improvement of both the reliability and bandwidth efficiency in wireless communication systems. This has spurred a notable recent thrust area of research and development, namely, the so-called Multiple-Input Multiple-Output (MIMO) technology, that is aimed at making better use of multiple antennas in wireless communications.; One of the important discoveries in MIMO communications is the existence of an optimal tradeoff curve between data rate (bandwidth efficiency) and performance (reliability). This and the growing diversification in data rates and quality of services of wireless communications call for MIMO transmission schemes that can provide various rate-performance tradeoffs. This dissertation is aimed at developing new MIMO transmission schemes with flexible rate-performance tradeoff.; For flat fading channels, a formal approach for the design of a class of general space-time block codes, i.e., linear dispersion codes, is developed. This results in a new multi-layered linear coding scheme. With a common encoding structure, various rate-performance tradeoffs can be achieved simply by choosing the number of data layers that enter the encoder. Once the data rate is chosen, phase shifts among input symbols can be applied to optimize the performance without loss of mutual information.; For frequency-selective MIMO channels, a new space-time orthogonal frequency division multiplexing (OFDM) modulation scheme is introduced to avoid the complicated code design. This is a significant departure from the conventional MIMO OFDM design. The benefit is that a large number of existing codes optimized for single-input fading channels are also optimal for MIMO channels. In addition, it allows an easy spatial multiplexing scheme to support multiple data layers and hence various data rates.; Last, a pilot-symbol-aided channel estimator is developed for the proposed MIMO OFDM transmission scheme. It is shown that, in terms of performance, the proposed scheme is equivalent to the conventional MMSE channel estimator but with much reduced computational complexity. |
