Thumbnail Image
Publication or External Link
Rong, Beiyu
Ephremides, Anthony
In wireless networks, a transmitted message may successfully reach multiple nodes simultaneously, which is referred to as the Wireless Multicast Advantage. As such, intermediate nodes have the ability to capture the message and then contribute to the communication toward the ultimate destination by cooperatively relaying the received message. This enables cooperative communication, which has been shown to counteract the effects of fading and attenuation in wireless networks. There has been a great deal of work addressing cooperative methods and their resulting benefits, but most of the work to date has focused on physical-layer techniques and on information-theoretic considerations. While compatible with these, the main thrust of this dissertation is to explore a new approach by implementing cooperation at the network layer. First, we illustrate the idea in a multi-hop multi-access wireless network, in which a set of source users generate packets to deliver to a common destination. An opportunistic and dynamic cooperation protocol is proposed at the network level, where users with a better channel to the destination have the capability and option to relay packets from users that are farther afield. The proposed mode of cooperation protocol is new and relies on MAC/Network-level of relaying, but also takes into account physical-layer parameters that determine successful reception at the destination and/or the relay. We explicitly characterize the stable throughput and average delay performance. Our analysis reveals that cooperation at the network layer leads to substantial performance gains for both performance metrics. Next, on top of the network-layer cooperation, we investigate enhanced cooperative techniques that exploit more sophisticated physical-layer properties. Specifically, we consider dynamic decode-and-forward, superposition coding, and multipacket reception capability, and we quantify the extent to which the enhancement techniques can further improve the stable throughput region. Then we revert back to the two-user multi-access channel with single-packet reception, which has been extensively studied in the case of no cooperation. After cooperation is permitted between the two users, we revisit the relationship between the <italic>stability region</italic> and the <italic>throughput region</italic> under both scheduled access and random access schemes. Finally, we shift our focus from the packet-level to bit-level multi-access channels. By exploiting the bit-nature of a packet, we create a bridge between traditional physical-layer-based transmission rates and classical MAC/Network-layer-based throughput rates. We first obtain the closed form of the stability region in bits/slot. Then, as a separate, but related issue, we look at the minimum delivery time policy; for any initial queue size vector, the optimal policy that empties all bits in the system within the shortest time is characterized.