Cross -layer interactions and optimizations in wireless networks

Wireless networks are characterized by the broadcast nature of the wireless channel, strong path loss, time varying fading and shadowing, and limited battery and processing power of the devices. These properties of the wireless physical layer interact in a complex manner with the higher layers of the protocol stack (MAC and routing), and present several interesting challenges when analyzing, dimensioning and designing wireless networks. The focus of this thesis is on studying these cross-layer interactions, and designing schemes which take advantage of these interactions. We study three different wireless networks, viz, wireless sensor networks, ad hoc networks, and cellular networks. In the context of sensor networks, we study the problem of guaranteeing a certain minimum network lifetime by optimally dimensioning the node battery energy, the extent of clustering in the network, and the communication range of the nodes. We also obtain design guidelines for choosing between multi-hop and single hop modes of communication, and between homogeneous and heterogeneous (multiple types of nodes) sensor networks. We also study the problem of determining bounds on the capacity of large ad hoc networks from a cross-layer perspective. Unlike previous work, our results are based on a realistic link layer model that takes into account the mapping between the Signal to Interference and Noise Ratio (SINR) and the packet error rate. Finally, in the context of cellular networks, we study the cross-layer effect of network load in the neighboring sectors of a terminal on the quality of its forward link. We analytically characterize this relationship, and propose a novel SINR estimation scheme that results in higher throughput, especially for terminals located near the cell boundary