Design of efficient decoders for wireless MIMO systems

Abstract

Demands for high data rate and high quality services in wireless communications, driven by wireless LANs, cellular mobile, and multimedia services, have been rapidly increasing throughout the world. However, the limited availability of spectrum and a complex time-varying physical channel are the main factors facing wireless designers. Therefore, in order to meet the demands, innovative techniques that have capability of improving spectral efficiency as well as transmission reliablity are desirable. The utilization of multiple antennas at the receiver and/or at the transmitter in a wireless system, widely known as wireless multiple-input multiple-output (MIMO) systems, is an emerging technology, which is expected to make high-data-rate and high-quality wireless communication systems a reality. Current transmission schemes over MIMO channels typically fall into three main categories: maximizing diversity (via the use of space-time coding), maximizing data rate (through the use of spatial multiplexing, V-BLAST), or increasing data rate and diversity simultaneously. In this thesis, we focus mainly on the design of quasi-orthogonal space-time block codes (QO-STBCs) and on the development of low complexity optimal and suboptimal $decoders^{1}$ for QO-STBCs and other space-time block codes as well. Specifically, we present the rate-one fulldiversity QO-STBCs for any number of transmit antennas. The proposed QO-STBCs are delay optimal for any even number of transmit antennas. As a result, in many situations, they are more advantageous than existing rate-one QO-STBCs in terms of detection complexity. In addition, several low-complexity maximum-likelihood (ML) decoders, namely, LC-ML, VLCMLDec1, VLCMLDec2, PMLD and QMLD, are devised for the detection of the proposed QO-STBCs as well as of other existing QO-STBCs. Then, based on the PMLD and QMLD, we construct two computationally efficient complex sphere decoders, called PSD and QSD. Finally, by combining the PSD and Q...