Wireless networks have traditionally been deployed with interference avoidance as the primary objective to ensure high throughput. Recently, a new and exciting paradigm has emerged in which interfering communications from different users are viewed as a way to spread "common information" across the network, thereby allowing for cooperation among otherwise uncoordinated transmitters. Cooperative communications holds the potential to enhance the communication capabilities of all users in the network simultaneously if users share their resources with their neighbors. This research aims at establish a theoretical foundation for collaborative communications by identifying the optimal resource sharing conduct, here referred to as etiquette, among users in peer-to-peer wireless networks. It is centered on the interference channel with generalized feedback, a novel information-theoretic network model that incorporates the knowledge a communicating pair of users can infer about other concurrent communications from information overheard. This research has two main thrusts: (1) developing innovative coding and signaling techniques for resource sharing to improve capacity, and investigate the associated scaling laws, (2) designing efficient and distributed multi-access and routing protocols to realize the benefits of collaborative communications with minimal overhead. The goal is to shed light on a richer and more powerful class of cooperation modes than currently proposed multi-hop relaying, and to lay the foundations for a deeper understanding of the ultimate tradeoffs involved in peer-to-peer networking. The technical approach includes comprehensive theoretical analysis and extensive computer simulations on real network data. In addition, a new multidisciplinary course that bridges Information Theory and Networking Theory is developed to expose students to the benefits of cross-fertilizations of the two disciplines to the design of collaborative communication networks.
The tremendous increase in wireless data traffic, since the introduction on the market of smart phones and tablets, forces communication engineers to rethink the cellar network paradigm. In this research we studied novel methods to increase the networks throughput, measured by the number of bits that can be communicated reliably from source(s) to receiver(s) per unit of time and per unit of frequency. In particular, we focused on allowing different devices to help one another in meeting their communication goals. This technology is termed "user cooperation" and includes as special cases relaying, feedback, and cognitive radio. In this research we focused on simple networks with two source-destination pairs that share the same wireless channel, thereby creating mutual interference to one another. We proposed communication strategies that are provably either exactly optimal or optimal to within a constant gap (i.e., a gap measures the distance between what is possibly and what is unattainable in terms of rates in the worst network scenario). These strategies for general cooperation were also specialized to the case of relaying, of feedback, and of cognitive radio. We also included practical constraints of low cost devices, such as the inability to receive and transmit at the same time. We compared the performance for different network topologies thereby identifying the conditions under which cooperation is most beneficial. Based on the insight gained from the two-user network, we extended some of the results to networks with an arbitrary number of users. The findings of this research are expected to impact the design of future broadband wireless networks: we identified startegies through which devices can better manage the spectrum resources (as comapred to what done in today's networks) thought mutual cooperation and intelligent interference management.
