Adaptive antenna arrays for third generation digital cellular systems

In this thesis we have applied the Eigen Spectral Decomposition Theorem to the signal covariance matrix used in the antenna array weight calculation. The cost functions investigated were MMSE and MSINR based. We have determined where the Reduced Rank techniques work well and where they brake down in performance. For the various wireless channel models investigated we have found the full rank, structured techniques introduced to work best. Here all the covariance matrix eigen-spectral information is maintained, but used to provide a matrix some structure. This structure can be accomplished by either forcing the eigenvalues to behave in a certain manner and/or using the channel estimates directly in the covariance matrix. This structure is actually missing in the conventional antenna array weights estimation techniques.; We have also exploited the eigen-spectra of the space-time covariance matrix for both the Symbol spaced and Fractionally spaced receivers. We have found performance improvement can be achieved by the use of forcing a particular structure to the estimated covariance matrix. We have also shown when the matrix rank is a useful parameter and when it is not. A few common space-time equalization techniques were investigated and compared against each other. We have found the architectures that were based on an oversampled covariance matrix performed better than those of a symbol rate sampled matrix.; Lastly, we have applied the above space-time equalization techniques to the 3G WCDMA cellular system. Here instead of using the dedicated pilot bits to estimate the array weights on the data channel, we have suggested a technique that allows us to use the data symbols directly to estimate the array weights. We have derived the finger output equations of the RAKE that enabled us to observe the fact that special eigen-spectral properties exist. We have shown significant performance improvement and discussed avenues where this improvement can be used such as cell range extension and/or capacity increase, etc. Especially with the WCDMA system proposing High Speed Downlink Packet Access (HSDPA) these techniques can be applicable for low spreading factor CDMA systems.