Random beamforming in heterogeneous networks.

摘要

With the progression of geometric beamforming techniques, the time to expand beyond isotropic networks has arrived. By taking into account the gains and beam patterns of individual elements in a beamforming array, a heterogeneous network may perform the same beamforming goals as an isotropic network. The focus of this work is the concept of randomly distributed arrays of nodes that wish to perform beamforming to maximize their radiation pattern towards an intended receiver while minimizing their pattern towards a set of interferers utilizing their position information. The use of the nodes' position information rather than the nodes' channel state information (CSI), called geometric beamforming, allows for the nodes to calculate their weights using significantly less data, and therefore less overhead in the cooperation phase prior to beamforming. The drawback is the relative ease of distributed calculations with the CSI model, where less information needs to be shared overall. In addition, utilizing position information rather than harvested CSI data allows for additional capabilities such as non-cooperative null steering. This work introduces a new heterogeneous beamforming model which takes anisotropic antenna patterns into account. Several applications of the new model are introduced and treated to both theoretical and numerical analysis including distributed weight calculations and initial research into tiered beamforming---a technique made possible by the new heterogeneous model.