Irregular and multi-channel sampling of operators (Q982556): Difference between revisions

The paper begins with a brief review of the classical sampling theorem for band-limited functions. Recall the Fourier transform on \(L^2({\mathbb{R}}^d)\) is densely defined by \[ {\mathfrak I}(f)(\xi)=\widehat f(\xi)= \int_{{\mathbb{R}}^d} f(t) e^{-2\pi it\cdot\xi}dt, \] where \(f\in L^1({\mathbb{R}}^d)\cap L^2({\mathbb{R}}^d)\). We take the domain of the operators to be defined as \[ S_\Omega= \prod^d_{k=1} \Biggl[-{\Omega_k\over 2},{\Omega_k\over 2}\Biggr] \] with \(\Omega= (\Omega_1,\Omega_2,\dots, \Omega_d)\). The Paley-Wiener space with bandwitdth \(\Omega\) is defined by \[ PW(S_\Omega)= \{f\in L^2({\mathbb{R}}^d):\text{ support }\widehat f\subseteq S_\Omega\}. \] The paper then continues with the raditional Nyqist-Landau sampling rate for a function in \(PW(S_\Omega) \)is defined to be the Lebesgue measure of the set, \(S_\Omega\). The following notation is defined to be \(A(F)<>B(F)\), \(F\in{\mathfrak I}\), if there exist positive constants \(c\) and \(C\ni cA(F)\leq B(F)\leq CA(F)\) for all objects \(F\) in the set \({\mathfrak I}\). The theorem that begins the paper is as follows: If we let \(\Omega= (\Omega_1,\Omega_2,\dots, \Omega_d)\), \(T= (T_1,T_2,\dots, T_d)\) with \(T_k\), \(\Omega_k> 0\) and \(T_k\cdot\Omega_k> 0\) for all \(k\in\mathbb{Z}\). For \({2\over T}= ({2\over T_1},\dots,{2\over T_d})\) and \(\varphi\in PW(S_{{2\over T}-\Omega})\) with \(\widehat\varphi= 1\) on \(S_\Omega\), then we have the sampling expansion, \[ f(t)= T\sum_{n\in\mathbb{Z}_d} f(nT)\varphi(t- nT),\;f\in PW(S_\Omega), \] with uniform and \(L^2({\mathbb{R}}^d)\) onvergence. Moreover, the sampling expansion above is stable in the sense that \[ \| f\|^2_{L^2}\asymp \sum_{n\in\mathbb{Z}^d} |f(nT)|^2,\;f\in PW(S_\Omega). \] The paper then begins the process using distributions. These classical sampling theorems for band-limited functions have recently been generalized to apply to so-called band-limited operators, that is, to operators with band-limited Kohn-Nirenberg symbols. Here, we discuss operator sampling versions of two of the most central extensions to the classical sampling theorem. In irregular operator sampling, the sampling set is not periodic with uniform distance. In multi-channel operator sampling, we obtain complete information on an operator by multiple operator sampling outputs.