Space-time coding and Beamforming are two main techniques utilized in multiple-input multiple-output (MIMO) wireless systems to increase... Show moreSpace-time coding and Beamforming are two main techniques utilized in multiple-input multiple-output (MIMO) wireless systems to increase channel capacity. In this work, we present new beamforming algorithms and space-time codes for MIMO wireless systems to reduce the complexity and enhance the performance. We present a new framework to obtain beamforming algorithms for code division multiple access (CDMA) and orthogonal frequency division multiplexing (OFDM) systems based on principal component analysis (PCA) approach. Four different PCA objective functions are used as examples to illustrate how to obtain new beamforming algorithms based on the proposed framework. Since two of the objective functions result in the same algorithm, three different beamforming algorithms are obtained for CDMA and OFDM systems. We show that the new algorithms offer a desirable performance while they have considerably lower computational complexity than the optimum algorithms. It is demonstrated that the proposed algorithms also have a fast convergence rate. We also propose new space-time codes for both coherent and non-coherent wireless systems which can provide lower symbol error rates than existing codes in the same class. For the coherent case, we propose a new quasi-orthogonal code to provide a full-rate full-diversity transmission. Unlike the other codes in the literature, the proposed code does not require the transmitted symbols to be chosen from two different constellations. Choosing all symbols from only one constellation reduces the required storage memory to half of that needed in other methods. It also avoids decreasing the minimum distance between the symbols and hence, avoids degrading the coding gain. For the non-coherent case, a new constellation design for differential unitary space-time modulation is proposed. We generate xi a set of unitary matrices from the multiplication of a simple unitary matrix with entries of 1 by diagonal matrices made from quaternary alphabets. Then, the proposed constellation is formed from the product of the optimum matrix chosen from this set and the optimized diagonal matrix first used for the cyclic codes. We show that the proposed constellation yields a larger diversity product as well as a lower union bound for pairwise error probability with the same or lower complexity compared to other existing constellations. PH.D in Electrical Engineering, May 2013 Show less
