Analysis and design of interference-limited wireless systems
- Author(s): Stamatiou, Konstantinos
- et al.
This thesis is concerned with the performance analysis and design of interference-limited wireless systems. Specifically, we consider wireless systems where transmissions from potential interferers are uncoordinated, i.e., there is no scheduling for the purpose of interference management. Special emphasis is placed on how the spatial distribution of the interferers affects the statistical characteristics of the interference power. The first topic we address is the performance analysis of the downlink of a cellular frequency-hopping multiple-access system, where coherent detection is employed. Channel estimation is performed with the help of hopping patterns which are broadcasted by the base-station and tracked by different users in the cell. In a time and frequency correlated channel, we identify which patterns result in the best performance in terms of the bit-error-probability at the receiver. We then consider the effect of the interference from the base-station of an adjacent cell and assess the importance of interference side information at the receiver on the bit-error-probability. The second topic we focus on is the analysis of random single-hop networks. The model we consider is that of a homogeneous Poisson point process of transmitters, each with a RX at a fixed distance. We evaluate the packet error probability and the network transmission capacity when the channel, consisting of fading and interference, is constant or varying during the packet transmission. The impact of coding, spatial diversity and multiple-stream transmission on the network capacity is identified with closed-form expressions in the small packet error probability regime. The third and final topic addressed is the evaluation of the mean end-to-end delay in random multi-hop networks. We extend the single-hop network model employed previously to accommodate routes, each consisting of a source, a number of relays and a final destination. Combining tools from the analysis of single-hop networks with queuing theory we derive a closed-form expression for the mean end-to-end delay over the typical route. The number of relays and their placement are determined such that the delay is minimized, for different packet arrival scenarios at the sources