Abstract

Impairments found in the wireless channel, such as destructive multipath fading, cannot be fully addressed by using coding alone, despite its recent advances. Recent results show that gains in capacity can be obtained by using multiple antenna elements at the transmitter and the receiver. Instead of mitigating the effects of the multipath fading, multiple-input multiple-output (MIMO) systems use the rich scattering channel to increase the capacity at no bandwidth cost. Methods include BLAST, which maximizes the rate, and space-time coding, which maximizes the diversity. In this thesis, the fundamental trade-off between rate and diversity is derived for binary codes. Simple codes that can perform at any realizable rate/diversity are designed. These codes are referred to as Space-Time Layered Block Codes (STLBC), since they are in effect a 1-dimensional code layered into a space-time code. By selecting speck 1-D codes, the required diversity can be achieved at the maximal allowable rate. In order to detect the new STLBC codes, an iterative MMSE detector is used jointly with a soft input soft-output decoder. This detector/decoder uses a principle similar to Turbo decoding. STLBC codes introduce the concept of coding between layers, which requires a detector capable of dealing with possibly dependent layers. The classical multiuser detector is modified to allow for dependence. Lastly, techniques to improve the performance of STLBC are presented. These are based on a new design criterion that ensures codebooks of high average rank, as well as making certain the information is spread out over the two available dimensions.