Exploiting spatial diversity for synchronization and decoding in MIMO communications

In recent years, there has been significant advances in broadband wireless access technologies that will revolutionize military, mobile internet and next-generation cellular systems. Physical layers designers now envision dramatic increases in spectral efficiency and coverage for next-generation systems. One of the most challenging hurdles is the dispersive non Line-Of-Sight (LOS) channel. Two solutions have emerged that turn channel multipath into a benefit. These are orthogonal frequency-division multiplexing (OFDM) and MIMO-based spatial multiplexing. Development of powerful new signal processing techniques are needed to realize the potential of these technologies. This thesis proposes two such receiver signal processing techniques. First proposed for OFDM systems is a novel carrier frequency offset (CFO) synchronization algorithm which exploits receiver spatial diversity. Second novel spatial multiplexed MIMO receiver structure is proposed.; Multi-antenna receivers, exploit diversity gain by coherent combining methods in order to recover data. CFO estimation does not benefit from coherent combining, since channel information is not available due to loss of OFDM subchannel orthogonality prior to CFO compensation. In this thesis, a Non-linear Least Squares (NLS) derived blind CFO estimator with a simple yet optimum receiver diversity combining scheme is proposed. Receive spatial diversity and OFDM subspace structure due the placement of virtual subchannels are exploited without required channel knowledge.; Spatial multiplexing potential for dramatic capacity gain is often difficult to realize, since channel conditions are often less than ideal, i.e. a degree of spatial correlation and co-channel interference can be expected. Current solutions overcome these impairments with added computational complexity and long training intervals. This thesis presents a solution based on the reduced-rank multi-stage Wiener filter (MSWF) first proposed by Goldstein and Reed for sensor array application. MSWF when combined with spatial multiplexing decoding method successive interference cancellation (SIC) is able realize significant signal subspace compression or rank-reduction. As a consequence of rank-reduction, MSWF converges much faster than full-rank methods thus permitting shorter training intervals and additional margin for dealing with channel impairments.