| The rapid growth of wireless communications makes it increasingly possible for people to communicate anywhere and anytime, and drives the need for technologies that improve link quality, cell coverage, spectrum usage, and capacity. Spatial and temporal signal processing is a key leverage to enhance the wireless network performance. The spatial dimension (antenna arrays) can improve spatial diversity to combat fading and offer spatial discrimination to suppress co-channel interference (CCI). Likewise the temporal dimension (filtering) provides equalization to combat inter-symbol interference (ISI) and enhances temporal diversity. In this thesis, we focus on the combined space-time signal processing at the receiver for time division multiple access (TDMA) cellular systems.; We begin with a discussion of wireless propagation and introduce the channel model. We then present the signal model and discuss some improvements in space-time processing gain through the spatial structure (antenna topology) and temporal structure (by exploiting the knowledge of the pulse shaping function). We then propose a new two stage space-time receiver architecture. The first stage is a multi-input-single-output (MISO) space-time linear filter designed to suppress co-channel interference by spatial-temporal discrimination without attacking the inter-symbol interference. The second stage is a scalar Viterbi equalizer designed to optimally handle the inter-symbol interference. This two stage space-time receiver treats co-channel and inter-symbol interference separately, and it is attractive since it uses a MMSE approach for the CCI which is optimal when the interference channel in not known, and uses an optimal Viterbi receiver for the ISI. We derive a joint optimization criterion to simultaneously determine the weight vector for the space-time linear filter, and to estimate the effective channel vector for the scalar Viterbi equalizer.; The remainder of the thesis deals with performance studies and comparative analyses of the proposed space-time receiver. The use of a multi-input-multi-output (MIMO) space-time linear filter to replace the MISO filter is discussed. We also demonstrate the consistency of the new receiver to the well known maximal-ratio-combiner and the optimum combiner in the narrowband scenario. Furthermore, the tradeoff between complexity and marginal performance improvement by using a whitening filter prior to the Viterbi equalizer is demonstrated analytically. Finally, the superior performance of this new space-time receiver is confirmed by simulation studies. |
