Enhancing spectrum utilization through cooperation and cognition in wireless systems

Abstract

We have seen a proliferation of wireless technologies and devices in recent years. The resulting explosion of wireless demand has put immense pressure on available spectrum. Improving spectrum utilization is therefore necessary to enable wireless networks to keep up with burgeoning demand. This dissertation presents a cognitive and cooperative wireless architecture that significantly enhances spectrum utilization. Specifically, it introduces four new systems that embody a cross-layer design for cognition and cooperation. The first system, SWIFT, is a cognitive cross technology solution that enables wideband devices to exploit higher layer network semantics to adaptively sense which portions of the spectrum are occupied by unknown narrowband devices, and weave the remaining unoccupied spectrum bands into a single high-throughput wideband link. Second, FARA is a cooperative system that enables multi-channel wireless solutions like 802.11 to dynamically use all available channels for all devices in a performance-aware manner by using information from the physical layer and allocating to each link the frequency bands that show the highest performance for that link. SourceSync, the third system, enables wireless nodes in last-hop and wireless mesh networks to cooperatively transmit synchronously in order to exploit channel diversity and increase reliability. Finally, MegaMIMO enables wireless throughput to scale linearly with the number of transmitters by enabling multiple wireless transmitters to transmit simultaneously in the same frequency bands to multiple wireless receivers without interfering with each other. The systems in this dissertation demonstrate the practicality of cognitive and cooperative wireless systems to enable spectrum sharing. Further, as part of these systems, we design several novel primitives - adaptive spectrum sensing, time alignment, frequency synchronization, and distributed phase-coherent transmission, that can serve as fundamental building blocks for wireless cognition and cooperation. Finally, we have implemented all four systems described in this dissertation, and evaluated them in wireless testbeds, demonstrating large gains in practice.