On the local integrability condition for generalised translation-invariant systems (Q2323313)

The frame properties of generalized shift-invariant systems in \(L^2(\mathbb{R}^n)\) have first been investigated by \textit{E. Hernández} et al. [J. Geom. Anal. 12, No. 4, 615--662 (2002; Zbl 1039.42032)] and \textit{A. Ron} and \textit{Z. Shen} [Constr. Approx. 22, No. 1, 1--45 (2005; Zbl 1080.42025)]. The term local integrability condition (LIC) was introduced in [Hernández et al., loc. cit.], but it also appears implicitly in [Ron and Shen, loc. cit.]. The importance of this concept is mainly due to the fact that it yields characterization results for frames in terms of ``simple'' equations. Loosely speaking, the LIC is a regularity condition that prohibits pathological counterexamples and without such a condition structural results that substantiate intuition might fail. There are weaker conditions than the LIC that still yields characterization results for frames; the weakest known condition is the so-called 1-unconditional convergence property (1-UCP) from [\textit{H. Führ} and \textit{J. Lemvig}, J. Funct. Anal. 276, No. 2, 563--601 (2019; Zbl 1407.42023)]. Both the LIC and the 1-UCP can be difficult to verify in practice as they depend delicately on both the generating functions \(g_{j,p}\) and the translation subgroups \(\Gamma_j\). Therefore, other less technical conditions are of interest. These conditions fall essentially in two classes as mainly conditions on the generating functions or on the translation subgroups: 1) ``weak'' integrability conditions on series involving the Fourier transform of generating functions \(g_{j,p}\), but not their translate (e.g., temperateness [Ron and Shen, loc. cit.] and Calderón integrability [\textit{M. Bownik} and \textit{J. Lemvig}, Trans. Am. Math. Soc. 363, No. 4, 1887--1924 (2011; Zbl 1219.42026)]), 2) counting estimates that bound the number of lattice points (stemming from the dual subgroups) that intersect with the translates of a compact set (e.g., roundedness [Ron and Shen, loc. cit.], lattice counting estimate [Bownik and Lemvig, loc. cit.]; [\textit{M. Bownik} and \textit{J. Lemvig}, Int. Math. Res. Not. 2017, No. 23, 7264--7291 (2017; Zbl 1405.42059)]; [\textit{C. K. Chui} et al., J. Fourier Anal. Appl. 8, No. 2, 173--200 (2002; Zbl 1005.42020)]; [\textit{K. Guo} and \textit{D. Labate}, Collect. Math. 57, No. 3, 295--307 (2006; Zbl 1136.42309)]; [Hernández et al., loc. cit.], Property X [\textit{D. Barbieri} et al., Appl. Comput. Harmon. Anal. 50, 326--352 (2021; Zbl 1464.42020)]). While roundedness is a rather strong assumption of the translation subgroups, the lattice counting estimate is much weaker and important in wavelet analysis [Bownik and Lemvig, loc. cit]. Property X is a variant of roundedness [Ron and Shen, loc. cit.] that is useful in the study of dilation-continuous wavelet systems [Barbieri et al., loc. cit.)]. The above ``less technical conditions'' have mainly been studied in \(L^2(\mathbb{R}^n)\) for either generalized shift-invariant systems or wavelet systems, and this paper extends the study to (continuous) generalized translation-invariant (GTI) systems in \(L^2(G)\), where \(G\) is a locally compact abelian group. Proposition~3.1 shows that the Calderón integrability condition, the temperateness condition and the uniform (lattice) counting estimate implies the LIC. Lemma 3.2 shows that LIC implies the Calderón integrability condition (cf. Proposition 2.7 in [\textit{M. Bownik} and \textit{J. Lemvig}, Trans. Am. Math. Soc. 363, No. 4, 1887--1924 (2011; Zbl 1219.42026)]). Example 3.4 is a key construction to show that the temperateness condition is not necessary for the Calderón integrability condition. Example~3.7 shows that the LIC and the local Calderón integrability condition can hold, while the uniform counting estimate fails. The last section of the paper restricts to special (locally compact abelian) groups. For compact abelian group and discrete abelian groups, the LIC is equivalent to, respectively, the Calderón integrability condition and the temperateness condition (Proposition 4.1 and 4.2). Theorem 4.4 generalizes a characterization of the LIC by \textit{G. Kutyniok} and \textit{D. Labate} [Colloq. Math. 106, No. 2, 197--220 (2006; Zbl 1113.43008)] to GTI systems on locally compact Abelian groups containing a compact open subgroup. Finally, wavelet systems in \(L^2(\mathbb{R}^n)\) are considered.