Improving spectrum utilization in wireless networks

With the proliferation of wireless-enabled devices, the information being transmitted on the wireless spectrum has increased manifolds. This explosive growth of wireless traffic has created spectrum crisis. Further, combined with the inherently lossy wireless medium, it is imperative to develop techniques that can significantly improve wireless spectrum efficiency. This dissertation develops three complementary techniques to enhance spectrum efficiency: (i) sending more information per transmission, (ii) sending more transmissions per spectrum, and (iii) selecting the right spectrum for transmission. More specifically, in (i), we observe that network coding allows us to send more information per transmission by combining (coding) multiple packets together in a single transmission and letting multiple receivers extract different information from the same transmission. However, wireless networks are inherently prone to loss and how to harness network coding gain under such conditions poses a significant challenge. To this end, we develop a novel routing protocol, called O3, which jointly optimizes network coding, opportunistic routing, and rate limiting. Multi-antenna devices (MIMO) dramatically increase wireless network capacity by sending multiple transmissions simultaneously. However, most existing work focuses on MIMO in single hop wireless networks, and how to effectively extend MIMO benefits to multihop wireless networks remain an open problem. In (ii), we propose a new routing protocol, called DM+, which is the first practical distributed MIMO routing protocol. It optimizes spatial multiplexing, routing, and rate limiting in the presence of interference. Using simulation and testbed experiments, we show it out-performs state-of-the-art shortest part routing and opportunistic routing protocols. Finally, in (iii), we examine spectrum selection at two different granularities: (a) selecting an appropriate channel to transmit a frame, and (b) selecting a subcarrier (within a channel) to transmit a symbol of the frame. In (a), we propose LBRH, a novel channel hopping algorithm that allows different nodes converge to a fair and efficient channel hopping sequence in a completely distributed fashion. In (b), we develop Smart-Fi, a series of techniques to harness the frequency diversity of the channel while transmitting the current frame. We demonstrate the effectiveness of both approaches using simulation and testbed experiments.