UC Santa Barbara

The potential performance gains from distributed Multiple-Input Multiple-Output (MIMO) have been established by theory, simulations and experimental demonstrations over the past decade. By synchronizing their signals in frequency and phase at the receiver, multiple cooperating transmitters achieve an N-fold increase in received power compared to noncoherent signal combination, or power pooling, and an N^2-fold increase in received power compared to that for a single transmitter. The first part of this thesis investigates algorithms for Distributed MIMO (DMIMO). A well-known one-bit feedback scheme is extended to wideband systems by introducing additional feedback bit to enforce phase continuity across OFDM subcarriers to utilize available system bandwidth. It is also shown, however, that low signal to noise ratio (SNR) is a fundamental limitation for the well-known one-bit feedback algorithm and its variants. This limitation motivates the development of explicit training strategies that can operate at low SNR while keeping the number of feedback bits low. Explicit training strategies are extended to wideband setting by simple interpolation over OFDM subcarriers and sparse time domain modeling of outdoor channels.

The second part of this thesis presents two concept systems that could be based around the DMIMO paradigm: distributed base station and distributed 911. The use of a distributed base station enables distributed transmit beamforming at large carrier wavelengths to achieve significant range extension and increased downlink data rate, providing a low-cost infrastructure for applications such as rural broadband. The proposed cross-layer design of distributed base station is based on explicit training with heavily quantized feedback. The feasibility of such a system is explored using outage capacity analysis for both the downlink and the feedback link. System performance is quantified in different channel conditions. The distributed 911 concept system enables multiple nodes to communicate with a distant search-and-rescue vehicle which tries to approach to the nodes by using radio frequency trail of the periodic beacons transmitted by the nodes. As an initial step to realize this, an algorithm using Doppler frequency measurements is devised for a quasi-stationary radio frequency emitting source sought by an unmanned aerial vehicle (UAV). Our algorithm uses periodic beacons from the emitter to continuously adapt its trajectory by using consecutive frequency measurements and provides a great improvement over prior received signal strength based source seeking algorithms.