Differential space-time modulation and unitary constellation design

Wireless communication systems with multiple antennas have better performance than single-antenna systems in a fading channel since the channel capacity increases with the number of antennas and the potential diversity order is the product of the numbers of transmit and receive antennas. This dissertation addresses the transmission for the situation when the channel state information is not available at the receiver and information is conveyed by a sequence of unitary matrices that perform differential space-time modulation (DSTM). Our objective is to design flexible unitary constellations for DSTM systems.; First, the importance of unitary matrix constellations is discussed from an overview of the fundamentals on DSTM. The design criteria for unitary codes are discussed based on the analysis of the parameterization of the unitary matrices. There are two directions for our design of unitary constellations.; The first design type is an extension of the block orthogonal code. The traditional unitary code from an orthogonal design uses PSK constellations. We consider using QAM constellations to construct unitary codes from orthogonal codes. By using QAM constellations, we can obtain more available transmission rates. The simulation results shows that the orthogonal codes with QAM have higher coding gain than the codes with PSK at the same transmission rate. Since block orthogonal codes can obtain full diversity, we study them for uncoded DSTM systems.; The second category of our unitary design is directed at obtaining unitary codes that are scalable for any number of transmit antennas, with a high transmission rate. We propose a design scheme where a unitary matrix is built with partitioned matrices. A recursive unitary design is then presented which constructs a unitary matrix in a recursive pattern. Since the recursive code cannot achieve full diversity, it is applied to coded DSTM, where DSTM is concatenated with serially concatenated convolutional codes. The full diversity is verified by simulation results.; In the iterative decoding for the concatenated DSTM, we propose a complexity reduction algorithm for the modulation decoder where the computation of the bit extrinsic information is based on a candidate list. With appropriate choices for list sizes in the decoder, the performance of the decoder with full complexity can be approached with low computational complexity.