Multiplexing, scheduling, and multicasting strategies for antenna arrays in wireless networks

Abstract : A transmitter antenna array has the ability to direct data simultaneously to multiple receivers within a wireless network, creating potential for a more integrated view of algorithmic system components. In this thesis, such a perspective informs the design of two system tasks: the scheduling of packets from a number of data streams into groups; and the subsequent spatial multiplexing and encoding of these groups using array processing. We demonstrate how good system designs can help these two tasks reinforce one another, or alternatively enable tradeoffs in complexity between the two. Moreover, scheduling and array processing each benefit from a further awareness of both the fading channel state and certain properties of the data, providing information about key flexibilities, constraints and goals. Our development focuses on techniques that lead to high performance even with very low-complexity receivers. We first consider spatial precoding under simple scheduling and propose several extensions for implementation, such as a unified time-domain precoder that compensates for both cross-channel and intersymbol interference. We then show how more sophisticated, channel-aware scheduling can reduce the complexity requirements of the array processing. The scheduling algorithms presented are based on the receivers' fading channel realizations and the delay tolerances of the data streams. Finally, we address the multicasting of common data streams in terms of opportunities for reduced redundancy as well as the conflicting objectives inherent in sending to multiple receivers. Our channel-aware extensions of space-time codes for multicasting gain several dB over traditional versions that do not incorporate channel knowledge.