Escalating demands for high performance wireless systems requires the confluence of smart communication methods, network protocols and ongoing advances in fabrication technologies, in order to bring smaller form factor mobile handsets to market. On par with these trends, this thesis focuses on two main areas, namely, Multiple Antenna Systems and Adaptive MAC Design to improve wireless system performance. The first part of this research work presents a mathematical framework for characterizing the performance of cellular mobile radio systems equipped with smart antennas at the mobile handset to suppress a few dominant cochannel interferers (CCI) out of a total of L active independent but non-identically distributed Rayleigh faded CCI signals. Earlier works on this subject chose an unrealistic i.i.d assumption for the cochannel interferers. Since the CCI signals are of dissimilar signal strengths in practical operating environments, the premise of i.n.d fading statistics for the cochannel interferers is more realistic. In the subsequent section an analytical framework to investigate the benefits of a hybrid antenna array using selective interference nulling (SIN) and maximal ratio combining (MRC) in mobile radio environments is developed. In the second part of this thesis, we explore the performance gains that can be achieved by exploiting the synergy resulting from the combination of the MAC and the physical layer of a wireless network. As in a traditional design, the physical layer is responsible for providing error protection for the transmission packets while the MAC layer allocates transmission bandwidth to the contending users. However, in the proposed scheme the MAC layer makes slot assignment decisions based on the channel state information (CSI) from the physical layer. / Master of Science

Managing interference is the main technical challenge in wireless networks. Multiple input multiple output (MIMO) methods are key components to overcome the interference bottleneck and deliver higher data rates. The most efficient MIMO techniques require channel state information (CSI). In practice, this information is inaccurate due to errors in CSI acquisition, as well as mobility and delay. CSI inaccuracy reduces the performance gains provided by MIMO. When compounded with uncoordinated intercell interference, the degradation in MIMO performance is accentuated. This dissertation investigates the impact of CSI inaccuracy on the performance of increasingly complex interference limited networks, starting with a single interferer scenario, continuing to a heterogeneous network with a femtocell overlay, and finishing with a clustered multicell coordination model for randomly deployed transmitting nodes. First, this dissertation analyzes limited feedback beamforming and precoded spatial multiplexing over temporally correlated channels. Assuming uncoordinated interference from one dominant interferer, using Markov chain convergence theory, the gain in the average successful throughput at the mobile user is shown to decrease exponentially with the feedback delay. The decay rate is amplified when the user is interference limited. Interference cancellation methods at the receiver are shown to mitigate the effect of interference. This work motivates the need for practical MIMO designs to overcome the adverse effects of interference. Second, limited feedback beamforming is analyzed on the downlink of a more realistic heterogeneous cellular network. Future generation cellular networks are expected to be heterogeneous, consisting of a mixture of macro base stations and low power nodes, to support the increasing user traffic capacity and reliability demand. Interference in heterogeneous environments cannot be coordinated using traditional interference mitigation techniques due to the on demand and random deployment of low power nodes such as femtocells. Using tools from stochastic geometry, the outage and average achievable rate of limited feedback MIMO is computed with same-tier and cross-tier interference, and feedback delay. A hybrid fixed and random network deployment model is used to analyze the performance in a fixed cell of interest. The maximum density of transmitting femtocells is derived as a function of the feedback rate and delay. The detrimental effect of same-tier interference is quantified, as the mobile user moves from the cell-center to the cell-edge. The third part of this dissertation considers limited coordination between randomly deployed transmitters. Building on the established degrading effect of uncoordinated interference on practical MIMO methods, and the analytical tractability of random deployment models, interference coordination is analyzed. Using multiple antennas at the transmitter for interference nulling in ad hoc networks is first shown to achieve MIMO gains using single antenna receivers. Clustered coordination is then investigated for cellular systems with randomly deployed base stations. As full coordination in the network is not feasible, a random clustering model is proposed where base stations located in the same cluster coordinate. The average achievable rate can be optimized as a function of the number of antennas to maximize the coordination gains. For multicell limited feedback, adaptive partitioning of feedback bits as a function of the signal and interference strength is proposed to minimize the loss in rate due to finite rate feedback. / text