English

Associating vectors in $\CC^n$ with rank 2 projections in $\RR^{2n}$: with applications

Functional Analysis 2017-03-09 v1

Abstract

We will see that vectors in \CCn\CC^n have natural analogs as rank 2 projections in \RR2n\RR^{2n} and that this association transfers many vector properties into properties of rank two projections on \RR2n\RR^{2n}. We believe that this association will answer many open problems in \CCn\CC^n where the corresponding problem in \RRn\RR^n has already been answered - and vice versa. As a application, we will see that phase retrieval (respectively, phase retrieval by projections) in \CCn\CC^n transfers to a variation of phase retrieval by rank 2 projections (respectively, phase retrieval by projections) on \RR2n\RR^{2n}. As a consequence, we will answer the open problem: Give the complex version of Edidin's Theorem \cite{E} which classifies when projections do phase retrieval in \RRn\RR^n. As another application we answer a longstanding open problem concerning fusion frames by showing that fusion frames in \CCn\CC^n associate with fusion frames in \RR2n\RR^{2n} with twice the dimension. As another application, we will show that a family of mutually unbiased bases in \CCn\CC^n has a natural analog as a family of mutually unbiased rank 2 projections in \RR2n\RR^{2n}. The importance here is that there are very few real mutually unbiased bases but now there are unlimited numbers of real mutually unbiased rank 2 projections to be used in their place. As another application, we will give a variaton of Edidin's theorem which gives a surprising classification of norm retrieval. Finally, we will show that equiangular and biangular frames in \CCn\CC^n have an analog as equiangular and biangular rank 2 projections in \RR2n\RR^{2n}.

Keywords

Cite

@article{arxiv.1703.02657,
  title  = {Associating vectors in $\CC^n$ with rank 2 projections in $\RR^{2n}$: with applications},
  author = {Peter G. Casazza and Desai Cheng},
  journal= {arXiv preprint arXiv:1703.02657},
  year   = {2017}
}