betaFeedback

WorksMathematics

Sobolev spaces arising from the Stockwell transform on Gelfand pairs

Abstract

We introduce a family of Sobolev-type spaces on Gelfand pairs, built from the Stockwell transform: a function belongs to the space of a given smoothness when its Stockwell transform, multiplied by a weight that grows with the smoothness, is square integrable. We prove that these spaces are Hilbert spaces, that they form a decreasing scale under continuous embeddings, and that an interpolation inequality controls the norm at an intermediate smoothness by the norms at the two endpoints. Under integrability conditions on the weight and on the family of analysing windows, they embed continuously into Lebesgue spaces. Under a local domination condition on the window, we show that their elements have bounded continuous representatives. On the Euclidean motion group, where the Stockwell transform reduces to a windowed Hankel transform, this gives boundedness and continuity as soon as the smoothness exceeds half the dimension, which is the classical threshold for Euclidean spaces.

2020 Mathematics Subject Classification: 43A90, 43A85, 46E35, 42C40

Introduction

The Stockwell transform, introduced by Stockwell, Mansinha and Lowe [12], can be viewed either as a short-time Fourier transform with a frequency-dependent window, or as a phase-corrected wavelet transform, and it has since become a standard tool for analysing non-stationary signals in seismology, biomedical imaging, and mechanical fault diagnosis. While the Stockwell transform was first developed on the real line and on Euclidean space, subsequent work has sought to place it within the broader framework of abstract harmonic analysis, in the same way the short-time Fourier and wavelet transforms have already been extended to locally compact groups and homogeneous spaces. Esmaeelzadeh carried out the first such extension [5, 6], defining the Stockwell transform, together with its associated localization operators, on locally compact abelian groups. More recently, this construction was pushed further, to the setting of Gelfand pairs [4], a class that strictly contains locally compact abelian groups and that includes many spaces of independent geometric interest, such as spheres, hyperbolic spaces, and other symmetric spaces. Gelfand pairs occupy a distinguished position in noncommutative harmonic analysis; although the ambient group need not be commutative, the algebra of bi-invariant functions still is, so that a spherical Fourier analysis remains available. Working at the level of Gelfand pairs therefore makes it possible to transport time-frequency methods, originally designed for commutative groups, to a much wider and geometrically richer class of settings.

On the other hand, Sobolev spaces have long played a central role in the study of partial differential equations and integral equations, providing the natural framework in which questions of regularity, boundedness, and embedding are addressed. The interplay between Sobolev-type smoothness and time-frequency transforms has proved fruitful. Classically, Sobolev spaces of Bessel potentiel type are built from the Fourier transform on Euclidean spaces [1] and on locally compact groups/hypergroups [7, 9] and homogeneous spaces [2, 8, 10].

The purpose of the present paper is to construct a family of Sobolev-type spaces associated with the Stockwell transform on Gelfand pairs, and to study some of their major properties. We show that these spaces carry a natural Hilbert space structure, we establish a scale of continuous embeddings relating spaces of different smoothness to one another and to the underlying square-integrable functions, and we obtain an embedding into Lebesgue spaces under a suitable integrability condition. We further show, under a local domination hypothesis, that sufficiently smooth elements of these spaces are in fact bounded and continuous. Throughout, the theory is illustrated on the Euclidean motion group, a noncommutative Gelfand pair on which it recovers, as a special case, the classical smoothness threshold for the continuous embedding into bounded continuous functions on Euclidean spaces.

The remainder of the paper is organized as follows. Section 2 collects the background material on harmonic analysis on Gelfand pairs and on the Stockwell transform needed in the sequel. Section 3 contains the main results.

Preliminaries

In this section, we provide the tools needed in the sequel.

Harmonic analysis on Gelfand pairs

For this subsection we refer to [3, 13]. Let GG be a locally compact group (assumed to be Hausdorff once for all) with a left Haar measure and let KK be a compact subgroup of GG endowed with its normalized Haar measure. A function f:G→Cf:G\to\mathbb{C} is said to be KK-bi-invariant if it satisfies the property:

∀x∈G, ∀k,k′∈K,\forall x\in G,\ \forall k,k'\in K,
f(kxk′)=f(x).f(kxk')=f(x).

Let Cc(G//K)C_c(G//K) and Lp(G//K)L^p(G//K), 1≤p<∞1\leq p<\infty, denote respectively the space of complex-valued continuous compactly supported KK-bi-invariant functions on GG and the space of complex-valued KK-bi-invariant Haar pp-integrable functions on GG. For 1≤p<∞1\leq p<\infty, Lp(G//K)L^p(G//K) is a complex Banach space with respect to the norm

∥f∥Lp=(∫G∣f(x)∣p)1p.\lVert f\rVert_{L^p}=\left(\int_G\lvert f(x)\rvert^p\right)^{\frac{1}{p}}.

The convolution of f,g∈L1(G//K)f,g\in L^1(G//K) is defined by

f∗g(x)=∫Gf(y)g(y−1x) dy,x∈G.f*g(x)=\int_G f(y)g(y^{-1}x)\,dy,\quad x\in G.

The pair (G,K)(G,K) is called a Gelfand pair if Cc(G//K)C_c(G//K) is a commutative convolution algebra; by density of Cc(G//K)C_c(G//K) in L1(G//K)L^1(G//K), this is equivalent to say that L1(G//K)L^1(G//K) is a commutative convolution algebra. It is well known that if (G,K)(G,K) is a Gelfand pair, then GG is unimodular [3 Proposition 6.1.2].

A continuous KK-bi-invariant function φ\varphi is called a spherical function for the Gelfand pair (G,K)(G,K) if the map χφ:Cc(G//K)→C\chi_\varphi: C_c(G//K) \to\mathbb{C}, f↦∫Gf(x)φ(x−1)dxf \mapsto\int_G f(x)\varphi(x^{-1})dx is a nontrivial character of the convolution algebra Cc(G//K)C_c(G//K), that is

χ(f∗g)=χ(f)χ(g),f,g∈Cc(G//K).\chi(f * g) = \chi(f)\chi(g), \qquad f,g \in C_c(G//K).

The set of bounded spherical functions for the Gelfand pair (G,K)(G,K) is denoted by Sb(G,K)S_b(G,K). The spherical Fourier transform for f∈L1(G//K)f \in L^1(G//K) is defined by

f^(φ)=∫Gf(x)φ(x−1)dx,φ∈Sb(G,K).\widehat{f}(\varphi) = \int_G f(x)\varphi(x^{-1})dx, \qquad\varphi\in S_b(G,K).

A function φ:G→C\varphi: G \to\mathbb{C} is said to be positive-definite function if

∑n=1N∑m=1Nφ(xn−1xm)zn‾zm≥0,∀N∈N, ∀x1,…,xN∈G, ∀z1,…,zN∈C.\sum_{n=1}^{N}\sum_{m=1}^{N}\varphi(x_n^{-1}x_m)\overline{z_n}z_m \ge0,\quad\forall N \in\mathbb{N},\ \forall x_1,\ldots,x_N \in G,\ \forall z_1,\ldots,z_N \in\mathbb{C}.

Theorem 2.1. [13 Proposition 8.4.2] Let φ\varphi be a positive-definite function. Then,

  1. ∀x∈G\forall x \in G, ∣φ(x)∣≤φ(e)|\varphi(x)| \le\varphi(e).

  2. ∀x∈G\forall x \in G, φ(x−1)=φ(x)‾\varphi(x^{-1}) = \overline{\varphi(x)}.

If φ\varphi is a spherical function, then φ(e)=1\varphi(e)=1 [3 Proposition 6.1.5].

Let S+S^+ denote the set of positive definite spherical functions for the Gelfand pair (G,K)(G,K). It carries a Plancherel measure which ensures the inversion formula for the spherical Fourier transform. Indeed, the spherical Fourier transform extends to an isometric isomorphism L2(G//K)⟶L2(S+)L^2(G//K) \longrightarrow L^2(S^+) and

f(x)=∫S+f^(φ)φ(x)dφ,f∈L2(G//K), x∈G.f(x) = \int_{S^+} \widehat{f}(\varphi)\varphi(x)d\varphi,\qquad f \in L^2(G//K),\ x \in G.

The Stockwell transform on Gelfand pairs

We recall from [4] the definition of the Stockwell transform on a Gelfand pair together with the two properties, the inversion formula and the Plancherel-type isometry, that will actually be used in the sequel.

Let (G,K)(G,K) be a Gelfand pair and α\alpha a topological automorphism of GG. There exists a positive constant δα\delta_\alpha such that d(α(x))=δαdxd(\alpha(x))=\delta_\alpha dx. It follows,

∫Gf(α(x))dx=δα−1∫Gf(x)dx,f∈Cc(G//K).\int_G f(\alpha(x))dx = \delta_\alpha^{-1}\int_G f(x)dx,\qquad f \in C_c(G//K).

Definition 2.2 ([4 Definition 1]). Let (G,K)(G,K) be a Gelfand pair, let α\alpha be a topological automorphism of GG and let θ∈L1(G//K)∩L2(G//K)\theta\in L^1(G//K)\cap L^2(G//K). The Stockwell transform of f∈L2(G//K)f \in L^2(G//K) with respect to α\alpha and θ\theta is

Sθ,αf(t,φ)=δα12∫Gf(x)φ(x)θ(α(t−1x))‾dx,t∈G, φ∈Sb(G,K).(1)S_{\theta,\alpha}f(t,\varphi)=\delta_\alpha^{\frac{1}{2}}\int_G f(x)\varphi(x)\overline{\theta(\alpha(t^{-1}x))}dx,\qquad t\in G,\ \varphi\in S_b(G,K). \tag*{(1)}

If we set

θα,φ,t(x)=φ(x)θ(α(t−1x)),\theta_{\alpha,\varphi,t}(x)=\varphi(x)\theta(\alpha(t^{-1}x)),

then formula (1) reads

Sθ,αf(t,φ)=δα12⟨f,θα,φ,t⟩.S_{\theta,\alpha}f(t,\varphi)=\delta_\alpha^{\frac{1}{2}}\langle f,\theta_{\alpha,\varphi,t}\rangle.

The space S+\mathcal{S}^{+} is endowed with the Radon measure dφ\mathrm{d}\varphi carrying the spherical Fourier inversion formula.

Set

L2(G×S+)={f:G×S+→C measurable such that ∫S+∫G∣f(t,φ)∣2 dtdφ<∞}.L^{2}(G \times\mathcal{S}^{+}) = \left\{ f : G \times\mathcal{S}^{+} \to\mathbb{C}\ \text{measurable such that}\ \int_{\mathcal{S}^{+}}\int_{G} \lvert f(t,\varphi)\rvert^{2}\,\mathrm{d}t\mathrm{d}\varphi< \infty\right\}.

equipped with the norm

∥f∥L2(G×S+)=(∫S+∫G∣f(t,φ)∣2 dtdφ)12.\lVert f\rVert_{L^{2}(G \times\mathcal{S}^{+})} = \left(\int_{\mathcal{S}^{+}}\int_{G} \lvert f(t,\varphi)\rvert^{2}\,\mathrm{d}t\mathrm{d}\varphi\right)^{\frac{1}{2}}.

Theorem 2.3 (Inversion formula, [4 Equation (12)]). Let θ∈L1(G//K)∩L2(G//K)\theta\in L^{1}(G//K) \cap L^{2}(G//K), θ≠0\theta\ne0. For every f∈L2(G//K)f \in L^{2}(G//K),

f(x)=δα12∥θ∥L22∫S+∫GSθ,αf(t,φ)θα,φ,t(x) dtdφ,x∈G.f(x) = \frac{\delta_{\alpha}^{\frac{1}{2}}}{\lVert\theta\rVert_{L^{2}}^{2}}\int_{\mathcal{S}^{+}}\int_{G} S_{\theta,\alpha}f(t,\varphi)\theta_{\alpha,\varphi,t}(x)\,\mathrm{d}t\mathrm{d}\varphi,\quad x \in G.

In particular, when ∥θ∥L2=1\lVert\theta\rVert_{L^{2}} = 1,

f(x)=δα12∫S+∫GSθ,αf(t,φ)θα,φ,t(x) dtdφ,x∈G.f(x) = \delta_{\alpha}^{\frac{1}{2}}\int_{\mathcal{S}^{+}}\int_{G} S_{\theta,\alpha}f(t,\varphi)\theta_{\alpha,\varphi,t}(x)\,\mathrm{d}t\mathrm{d}\varphi,\quad x \in G.

Theorem 2.4 (Plancherel-type identity, [4 Theorem 7]). Let (G,K)(G,K) be a Gelfand pair and let θ,η,f,g∈L2(G//K)\theta,\eta,f,g \in L^{2}(G//K). Then

⟨Sθ,αf,Sη,αg⟩L2(G×S+)=⟨f,g⟩L2(G//K)⟨η,θ⟩L2(G//K).\langle S_{\theta,\alpha}f,S_{\eta,\alpha}g\rangle_{L^{2}(G \times\mathcal{S}^{+})} = \langle f,g\rangle_{L^{2}(G//K)}\langle\eta,\theta\rangle_{L^{2}(G//K)}.

In particular, if moreover ∥θ∥L2=1\lVert\theta\rVert_{L^{2}} = 1, then Sθ,αS_{\theta,\alpha} is an isometry:

∥Sθ,αf∥L2(G×S+)=∥f∥L2,∀f∈L2(G//K).\lVert S_{\theta,\alpha}f\rVert_{L^{2}(G \times\mathcal{S}^{+})} = \lVert f\rVert_{L^{2}},\quad\forall f \in L^{2}(G//K).

Main results

We state the main definition of this article.

Definition 3.1. Let γ:G×S+→R+\gamma: G \times\mathcal{S}^{+} \to\mathbb{R}_{+} be a positive measurable function and s∈R+s \in\mathbb{R}_{+}. The Sobolev space associated with the Stockwell transform on the Gelfand pair (G,K)(G,K) is the set

Hγ,θ,αs(G//K)={f∈L2(G//K):∫S+∫G(1+γ(t,φ)2)s∣Sθ,αf(t,φ)∣2 dtdφ<∞}H^{s}_{\gamma,\theta,\alpha}(G//K) = \left\{ f \in L^{2}(G//K) : \int_{\mathcal{S}^{+}}\int_{G}(1+\gamma(t,\varphi)^{2})^{s}\lvert S_{\theta,\alpha}f(t,\varphi)\rvert^{2}\,\mathrm{d}t\mathrm{d}\varphi< \infty\right\}

endowed with the norm

∥f∥Hγ,θ,αs=(∫S+∫G(1+γ(t,φ)2)s∣Sθ,αf(t,φ)∣2 dtdφ)12.\lVert f\rVert_{H^{s}_{\gamma,\theta,\alpha}} = \left(\int_{\mathcal{S}^{+}}\int_{G}(1+\gamma(t,\varphi)^{2})^{s}\lvert S_{\theta,\alpha}f(t,\varphi)\rvert^{2}\,\mathrm{d}t\mathrm{d}\varphi\right)^{\frac{1}{2}}.

We first show that the Sobolev-type space just introduced carries a Hilbert space structure.

Theorem 3.2. Let (G,K)(G,K) be a Gelfand pair and let θ∈L1(G//K)∩L2(G//K)\theta\in L^{1}(G//K) \cap L^{2}(G//K) be such that ∥θ∥L2=1\lVert\theta\rVert_{L^{2}} = 1. The space Hγ,θ,αs(G//K)H^{s}_{\gamma,\theta,\alpha}(G//K) is a Hilbert space.

Proof. Define Γ:Hγ,θ,αs(G//K)→L2(G×S+)\Gamma:H^s_{\gamma,\theta,\alpha}(G//K)\to L^2(G\times S^+) by Γf=(1+γ(⋅,⋅)2)s2Sθ,αf(⋅,⋅)\Gamma f=(1+\gamma(\cdot,\cdot)^2)^{\frac{s}{2}}S_{\theta,\alpha}f(\cdot,\cdot). The linearity of Γ\Gamma is obvious. Furthermore, we have

∥f∥Hγ,θ,αs2=∫S+∫G(1+γ(t,φ)2)s∣Sθ,αf(t,φ)∣2 dt dφ=∥(1+γ(⋅,⋅)2)s2Sθ,αf(⋅,⋅)∥L2(G×S+)2=∥Γf∥L2(G×S+)2.\begin{aligned} \|f\|_{H^s_{\gamma,\theta,\alpha}}^2 &=\int_{S^+}\int_G(1+\gamma(t,\varphi)^2)^s|S_{\theta,\alpha}f(t,\varphi)|^2\,dt\,d\varphi\\ &=\|(1+\gamma(\cdot,\cdot)^2)^{\frac{s}{2}}S_{\theta,\alpha}f(\cdot,\cdot)\|_{L^2(G\times S^+)}^2\\ &=\|\Gamma f\|_{L^2(G\times S^+)}^2. \end{aligned}

So Γ\Gamma is a linear isometry, hence injective. In particular ∥⋅∥Hγ,θ,αs(G//K)\|\cdot\|_{H^s_{\gamma,\theta,\alpha}(G//K)} is the norm on Hγ,θ,αs(G//K)H^s_{\gamma,\theta,\alpha}(G//K), induced by the inner product

⟨f,h⟩Hγ,θ,αs=∫S+∫G(1+γ(t,φ)2)sSθ,αf(t,φ)Sθ,αh(t,φ)‾ dt dφ.\langle f,h\rangle_{H^s_{\gamma,\theta,\alpha}}=\int_{S^+}\int_G(1+\gamma(t,\varphi)^2)^sS_{\theta,\alpha}f(t,\varphi)\overline{S_{\theta,\alpha}h(t,\varphi)}\,dt\,d\varphi.

Let (fn)n(f_n)_n be a Cauchy sequence in Hγ,θ,αs(G//K)H^s_{\gamma,\theta,\alpha}(G//K). Since s≥0s\geq0, we have (1+γ(t,φ)2)s≥1(1+\gamma(t,\varphi)^2)^s\geq1 for all (t,φ)(t,\varphi), so by Theorem 2.4,

∥fn−fm∥L22=∥Sθ,α(fn−fm)∥L2(G×S+)2≤∥fn−fm∥Hγ,θ,αs2.\|f_n-f_m\|_{L^2}^2=\|S_{\theta,\alpha}(f_n-f_m)\|_{L^2(G\times S^+)}^2\leq\|f_n-f_m\|_{H^s_{\gamma,\theta,\alpha}}^2.

Thus, (fn)n(f_n)_n is also Cauchy in L2(G//K)L^2(G//K). Since L2(G//K)L^2(G//K) is complete, (fn)n(f_n)_n converges to some ff in L2(G//K)L^2(G//K).

On the other hand, (Γfn)n=((1+γ2)s2Sθ,αfn)n(\Gamma f_n)_n=((1+\gamma^2)^{\frac{s}{2}}S_{\theta,\alpha}f_n)_n is Cauchy in L2(G×S+)L^2(G\times S^+) as Γ\Gamma is an isometry, hence converges to some g∈L2(G×S+)g\in L^2(G\times S^+). That is, (Sθ,αfn)n(S_{\theta,\alpha}f_n)_n converges to (1+γ2)−s2g(1+\gamma^2)^{-\frac{s}{2}}g.

Since Sθ,αS_{\theta,\alpha} is an isometry L2(G//K)→L2(G×S+)L^2(G//K)\to L^2(G\times S^+), it is in particular continuous, so Sθ,αfnS_{\theta,\alpha}f_n converges to Sθ,αfS_{\theta,\alpha}f in L2(G×S+)L^2(G\times S^+). By uniqueness of the limit, Sθ,αf=(1+γ2)−s2gS_{\theta,\alpha}f=(1+\gamma^2)^{-\frac{s}{2}}g a.e., that is g=Γfg=\Gamma f.

Since Γf=g∈L2(G×S+)\Gamma f=g\in L^2(G\times S^+), f∈Hγ,θ,αs(G//K)f\in H^s_{\gamma,\theta,\alpha}(G//K) and

∥fn−f∥Hγ,θ,αs=∥Γfn−Γf∥L2(G×S+)=∥Γfn−g∥L2(G×S+)⟶0,\|f_n-f\|_{H^s_{\gamma,\theta,\alpha}}=\|\Gamma f_n-\Gamma f\|_{L^2(G\times S^+)}=\|\Gamma f_n-g\|_{L^2(G\times S^+)}\longrightarrow0,

as nn goes to ∞\infty. Hence, every Cauchy sequence in Hγ,θ,αs(G//K)H^s_{\gamma,\theta,\alpha}(G//K) converges in Hγ,θ,αs(G//K)H^s_{\gamma,\theta,\alpha}(G//K), which is therefore a Hilbert space. □

Next, we show that increasing the smoothness parameter yields a decreasing scale of spaces, in the sense of continuous embeddings.

Theorem 3.3. Let (G,K)(G,K) be a Gelfand pair and let θ∈L1(G//K)∩L2(G//K)\theta\in L^1(G//K)\cap L^2(G//K) be such that ∥θ∥L2=1\|\theta\|_{L^2}=1. If τ>σ≥0\tau>\sigma\geq0, then the space Hγ,θ,ατ(G//K)H^\tau_{\gamma,\theta,\alpha}(G//K) is continuously embedded into Hγ,θ,ασ(G//K)H^\sigma_{\gamma,\theta,\alpha}(G//K) with

∥f∥Hγ,θ,ασ≤∥f∥Hγ,θ,ατ,∀f∈Hγ,θ,ατ(G//K).\|f\|_{H^\sigma_{\gamma,\theta,\alpha}}\leq\|f\|_{H^\tau_{\gamma,\theta,\alpha}},\quad\forall f\in H^\tau_{\gamma,\theta,\alpha}(G//K).

Proof. Since (1+γ(t,φ)2)≥1(1+\gamma(t,\varphi)^2)\geq1, then for τ>σ≥0\tau>\sigma\geq0, we have (1+γ(t,φ)2)τ≥(1+γ(t,φ)2)σ(1+\gamma(t,\varphi)^2)^\tau\geq(1+\gamma(t,\varphi)^2)^\sigma. The assertion follows. □

We next establish an interpolation inequality, showing that the norm at an intermediate smoothness level is controlled by the norms at the two endpoint levels.

Theorem 3.4. Let s1,s2s_1,s_2 be nonnegative real numbers satisfying s1<s2s_1<s_2. Let ζ∈[0,1]\zeta\in[0,1], put s=(1−ζ)s1+ζs2s=(1-\zeta)s_1+\zeta s_2. For all f∈Hγ,θ,αs(G//K)f\in H^s_{\gamma,\theta,\alpha}(G//K), we have

∥f∥Hγ,θ,αs≤∥f∥Hγ,θ,αs11−ζ∥f∥Hγ,θ,αs2ζ.\lVert f\rVert_{H^s_{\gamma,\theta,\alpha}}\leq\lVert f\rVert_{H^{s_1}_{\gamma,\theta,\alpha}}^{1-\zeta}\lVert f\rVert_{H^{s_2}_{\gamma,\theta,\alpha}}^\zeta.

Proof. Let ∀f∈Hγ,θ,αs(G//K)\forall f\in H^s_{\gamma,\theta,\alpha}(G//K). We have

∥f∥Hγ,θ,αs2=∫S+∫G(1+γ(t,φ)2)s∣Sθ,αf(t,φ)∣2 dt dφ=∫S+∫G(1+γ(t,φ)2)[(1−ζ)s1+ζs2]∣Sθ,αf(t,φ)∣2ζ+2(1−ζ) dt dφ=∫S+∫G(1+γ(t,φ)2)(1−ζ)s1(1+γ(t,φ)2)ζs2∣Sθ,αf(t,φ)∣2ζ∣Sθ,αf(t,φ)∣2(1−ζ) dt dφ≤(∫S+∫G(1+γ(t,φ)2)s1∣Sθ,αf(t,φ)∣2 dt dφ)1−ζ×(∫S+∫G(1+γ(t,φ)2)s2∣Sθ,αf(t,φ)∣2 dt dφ)ζ≤∥f∥Hγ,θ,αs12(1−ζ)∥f∥Hγ,θ,αs22ζ.\begin{aligned} \lVert f\rVert_{H^s_{\gamma,\theta,\alpha}}^2 &=\int_{S_+}\int_G(1+\gamma(t,\varphi)^2)^s\lvert S_{\theta,\alpha}f(t,\varphi)\rvert^2\,dt\,d\varphi\\ &=\int_{S_+}\int_G(1+\gamma(t,\varphi)^2)^{[(1-\zeta)s_1+\zeta s_2]}\lvert S_{\theta,\alpha}f(t,\varphi)\rvert^{2\zeta+2(1-\zeta)}\,dt\,d\varphi\\ &=\int_{S_+}\int_G(1+\gamma(t,\varphi)^2)^{(1-\zeta)s_1}(1+\gamma(t,\varphi)^2)^{\zeta s_2}\lvert S_{\theta,\alpha}f(t,\varphi)\rvert^{2\zeta}\lvert S_{\theta,\alpha}f(t,\varphi)\rvert^{2(1-\zeta)}\,dt\,d\varphi\\ &\leq\left(\int_{S_+}\int_G(1+\gamma(t,\varphi)^2)^{s_1}\lvert S_{\theta,\alpha}f(t,\varphi)\rvert^2\,dt\,d\varphi\right)^{1-\zeta}\\ &\qquad\times\left(\int_{S_+}\int_G(1+\gamma(t,\varphi)^2)^{s_2}\lvert S_{\theta,\alpha}f(t,\varphi)\rvert^2\,dt\,d\varphi\right)^\zeta\\ &\leq\lVert f\rVert_{H^{s_1}_{\gamma,\theta,\alpha}}^{2(1-\zeta)}\lVert f\rVert_{H^{s_2}_{\gamma,\theta,\alpha}}^{2\zeta}. \end{aligned}

We deduce

∥f∥Hγ,θ,αs≤∥f∥Hγ,θ,αs11−ζ∥f∥Hγ,θ,αs2ζ.\lVert f\rVert_{H^s_{\gamma,\theta,\alpha}}\leq\lVert f\rVert_{H^{s_1}_{\gamma,\theta,\alpha}}^{1-\zeta}\lVert f\rVert_{H^{s_2}_{\gamma,\theta,\alpha}}^\zeta.

In order to obtain embeddings into other Lebesgue spaces, we first establish the following auxiliary estimate, which bounds the Lebesgue norm of a function by a Lebesgue-type norm of its Stockwell transform.

Lemma 3.5. Let (G,K)(G,K) be a Gelfand pair and let θ∈L1(G//K)∩L2(G//K)\theta\in L^1(G//K)\cap L^2(G//K) with ∥θ∥L2=1\lVert\theta\rVert_{L^2}=1. Let 1≤p<∞1\leq p<\infty and 1≤q≤∞1\leq q\leq\infty, with conjugate exponent q′q', and set

Np,q=(∫S+∫G∥θα,φ,t∥Lp(G)q′ dt dφ)1q′N_{p,q}=\left(\int_{S_+}\int_G\lVert\theta_{\alpha,\varphi,t}\rVert_{L^p(G)}^{q'}\,dt\,d\varphi\right)^{\frac{1}{q'}}

(an essential supremum over (t,φ)(t,\varphi) if q′=∞q'=\infty). If Np,q<∞N_{p,q}<\infty, then for every f∈L2(G//K)f\in L^2(G//K),

∥f∥Lp≤δα12Np,q∥Sθ,αf∥Lq(G×S+).\lVert f\rVert_{L^p}\leq\delta_\alpha^{\frac12}N_{p,q}\lVert S_{\theta,\alpha}f\rVert_{L^q(G\times S_+)}.

Proof. Let g∈Cc(G//K)⊂L2(G//K)g\in C_c(G//K)\subset L^2(G//K). Since Sθ,αg(t,φ)=δα1/2⟨g,θα,φ,t⟩S_{\theta,\alpha}g(t,\varphi)=\delta_\alpha^{1/2}\langle g,\theta_{\alpha,\varphi,t}\rangle, Theorem 2.4 (with η=θ\eta=\theta, $\lVert\theta\rVert_{L^2}=1) gives

⟨f,g⟩=⟨Sθ,αf,Sθ,αg⟩=δα12∫S+∫GSθ,αf(t,φ)⟨θα,φ,t,g⟩ dt dφ.\langle f,g\rangle=\langle S_{\theta,\alpha}f,S_{\theta,\alpha}g\rangle=\delta_\alpha^{\frac12}\int_{S_+}\int_G S_{\theta,\alpha}f(t,\varphi)\langle\theta_{\alpha,\varphi,t},g\rangle\,dt\,d\varphi.

By Hölder’s inequality in GG, ∣⟨θα,φ,t,g⟩∣≤∥θα,φ,t∥Lp∥g∥Lp′|\langle\theta_{\alpha,\varphi,t},g\rangle| \le\|\theta_{\alpha,\varphi,t}\|_{L^p}\|g\|_{L^{p'}}, where p′p' is the conjugate exponent of pp. Then Hölder’s inequality in (t,φ)(t,\varphi) yields

∣⟨f,g⟩∣≤δα12∥g∥Lp′∫S+∫G∣Sθ,αf∣∥θα,φ,t∥Lp dt dφ≤δα12Np,q∥Sθ,αf∥Lq∥g∥Lp′.|\langle f,g\rangle| \le\delta_{\alpha}^{\frac{1}{2}}\|g\|_{L^{p'}}\int_{S^+}\int_G |S_{\theta,\alpha}f|\|\theta_{\alpha,\varphi,t}\|_{L^p}\,dt\,d\varphi\le\delta_{\alpha}^{\frac{1}{2}}N_{p,q}\|S_{\theta,\alpha}f\|_{L^q}\|g\|_{L^{p'}}.

Since ff is KK-bi-invariant, we have

∥f∥Lp=sup⁡{∣⟨f,g⟩∣:g∈Cc(G//K), ∥g∥Lp′≤1}\|f\|_{L^p}=\sup\{|\langle f,g\rangle|:g\in C_c(G//K),\ \|g\|_{L^{p'}}\le1\}

(valid for 1≤p<∞1\le p<\infty). Therefore,

∥f∥Lp≤δα12Np,q∥Sθ,αf∥Lq(G×S+).\|f\|_{L^p}\le\delta_{\alpha}^{\frac{1}{2}}N_{p,q}\|S_{\theta,\alpha}f\|_{L^q(G\times S^+)}.

□\square

Building on the previous lemma, we now show that a suitable condition on the weight yields a continuous embedding of the Sobolev-type space into a Lebesgue space.

Theorem 3.6. Let (G,K)(G,K) be a Gelfand pair and let θ∈L1(G//K)∩L2(G//K)\theta\in L^1(G//K)\cap L^2(G//K) with ∥θ∥L2=1\|\theta\|_{L^2}=1. Let s>0s>0, β>2\beta>2 and set p=2ββ+2∈(1,2)p=\frac{2\beta}{\beta+2}\in(1,2), so that 1p=12+1β\frac{1}{p}=\frac{1}{2}+\frac{1}{\beta}. Assume that (1+γ2)−s2∈Lβ(G×S+)(1+\gamma^2)^{-\frac{s}{2}}\in L^\beta(G\times S^+) and Np:=Np,p<∞N_p:=N_{p,p}<\infty (with Np,pN_{p,p} as in Lemma 3.5, i.e. with q=pq=p). Then Hγ,θ,αs(G//K)H^s_{\gamma,\theta,\alpha}(G//K) is continuously embedded in Lp(G//K)L^p(G//K) and

∥f∥Lp≤δα12Np∥(1+γ2)−s2∥Lβ(G×S+)∥f∥Hγ,θ,αs,f∈Hγ,θ,αs(G//K).\|f\|_{L^p}\le\delta_{\alpha}^{\frac{1}{2}}N_p\|(1+\gamma^2)^{-\frac{s}{2}}\|_{L^\beta(G\times S^+)}\|f\|_{H^s_{\gamma,\theta,\alpha}},\qquad f\in H^s_{\gamma,\theta,\alpha}(G//K).

Proof. Since 1p=12+1β\frac{1}{p}=\frac{1}{2}+\frac{1}{\beta}, the Hölder inequality gives

∥Sθ,αf∥Lp(G×S+)≤∥(1+γ2)s/2Sθ,αf∥L2(G×S+)∥(1+γ2)−s/2∥Lβ(G×S+)=∥f∥Hγ,θ,αs∥(1+γ2)−s/2∥Lβ.\begin{aligned} \|S_{\theta,\alpha}f\|_{L^p(G\times S^+)}\le\|(1+\gamma^2)^{s/2}S_{\theta,\alpha}f\|_{L^2(G\times S^+)}\|(1+\gamma^2)^{-s/2}\|_{L^\beta(G\times S^+)} \\ &=\|f\|_{H^s_{\gamma,\theta,\alpha}}\|(1+\gamma^2)^{-s/2}\|_{L^\beta}. \end{aligned}

Lemma 3.5 with q=pq=p concludes ∥f∥Lp≤δα12Np∥Sθ,αf∥Lp\|f\|_{L^p}\le\delta_{\alpha}^{\frac{1}{2}}N_p\|S_{\theta,\alpha}f\|_{L^p}.

□\square

We now give a second embedding into Lebesgue spaces, based on the endpoint case q=1q=1 of Lemma 3.5. It requires only that the window belongs to LpL^p and that the inverse of the weight is square integrable, and it involves no condition on the family of analysing windows.

Proposition 3.7. Let θ∈L1(G//K)∩L2(G//K)∩Lp(G//K)\theta\in L^1(G//K)\cap L^2(G//K)\cap L^p(G//K) with ∥θ∥L2=1\|\theta\|_{L^2}=1, 1≤p<∞1\le p<\infty, s>0s>0, and assume (1+γ2)−s/2∈L2(G×S+)(1+\gamma^2)^{-s/2}\in L^2(G\times S^+). Then Hγ,θ,αs(G//K)↪Lp(G//K)H^s_{\gamma,\theta,\alpha}(G//K)\hookrightarrow L^p(G//K) with

∥f∥Lp≤δα12−1p∥θ∥Lp∥(1+γ2)−s/2∥L2(G×S+)∥f∥Hγ,θ,αs.\|f\|_{L^p}\le\delta_{\alpha}^{\frac{1}{2}-\frac{1}{p}}\|\theta\|_{L^p}\|(1+\gamma^2)^{-s/2}\|_{L^2(G\times S^+)}\|f\|_{H^s_{\gamma,\theta,\alpha}}.

Proof. Since ∣φ∣≤1|\varphi|\le1 and

∫G∣θ(α(t−1x))∣p dx=δα−1∥θ∥Lpp\int_G|\theta(\alpha(t^{-1}x))|^p\,dx=\delta_{\alpha}^{-1}\|\theta\|_{L^p}^p

(left invariance of the Haar measure and the change of variables x↦α(x)x\mapsto\alpha(x)), we get Np,1≤δα−1/p∥θ∥LpN_{p,1}\le\delta_{\alpha}^{-1/p}\|\theta\|_{L^p}. Moreover, by Cauchy–Schwarz,

∥Sθ,αf∥L1≤∥(1+γ2)s/2Sθ,αf∥L2∥(1+γ2)−s/2∥L2.\|S_{\theta,\alpha}f\|_{L^1}\le\|(1+\gamma^2)^{s/2}S_{\theta,\alpha}f\|_{L^2}\|(1+\gamma^2)^{-s/2}\|_{L^2}.

Finally, apply Lemma 3.5 with q=1q=1, q′=∞q^{\prime}=\infty yields the result. □

We finally turn to pointwise regularity. The following theorem states that, under a local domination condition on the window, the elements of the Sobolev-type space are continuous, and that the space embeds continuously into the bounded continuous functions.

Theorem 3.8. Let (G,K)(G,K) be a Gelfand pair with GG first countable, and let α\alpha be a topological automorphism of GG with α(K)=K\alpha(K)=K. Let θ∈L1(G//K)∩L2(G//K)\theta\in L^{1}(G//K)\cap L^{2}(G//K) be continuous with ∥θ∥L2=1\|\theta\|_{L^{2}}=1, and let s≥0s\geq0. Consider the following hypotheses:

(H) There exist a KK-bi-invariant open neighbourhood VV of x0∈Gx_{0}\in G and h∈L1(G×S+)h\in L^{1}(G\times S^{+}) such that, for every y∈Vy\in V and for a.e. (t,φ)(t,\varphi),

(1+γ(t,φ)2)−s∣θ(α(t−1y))∣2≤h(t,φ).(1+\gamma(t,\varphi)^{2})^{-s}\left|\theta\left(\alpha(t^{-1}y)\right)\right|^{2}\leq h(t,\varphi).

(M) M:=sup⁡y∈G∫S+∫G(1+γ(t,φ)2)−s∣θ(α(t−1y))∣2 dt dφ<∞M:=\sup_{y\in G}\int_{S^{+}}\int_{G}(1+\gamma(t,\varphi)^{2})^{-s}\left|\theta\left(\alpha(t^{-1}y)\right)\right|^{2}\,dt\,d\varphi<\infty.

Then,

(a) If (H) holds at x0x_{0}, then for every f∈Hγ,θ,αs(G//K)f\in H^{s}_{\gamma,\theta,\alpha}(G//K) the integral

F(x)=δα12∫S+∫GSθ,αf(t,φ)φ(x)θ(α(t−1x)) dt dφF(x)=\delta_{\alpha}^{\frac{1}{2}}\int_{S^{+}}\int_{G}S_{\theta,\alpha}f(t,\varphi)\varphi(x)\theta\left(\alpha(t^{-1}x)\right)\,dt\,d\varphi

converges absolutely for x∈Vx\in V, FF is continuous on VV, and F=fF=f almost everywhere on VV. In particular ff has a representative which is continuous at x0x_{0}.

(b) If (H) holds at every point of GG and (M) holds, then every f∈Hγ,θ,αs(G//K)f\in H^{s}_{\gamma,\theta,\alpha}(G//K) has a representative F∈Cb(G//K)F\in C_{b}(G//K), and

∥F∥∞≤δα12M12∥f∥Hγ,θ,αs.\|F\|_{\infty}\leq\delta_{\alpha}^{\frac{1}{2}}M^{\frac{1}{2}}\|f\|_{H^{s}_{\gamma,\theta,\alpha}}.

Thus Hγ,θ,αs(G//K)↪Cb(G//K)H^{s}_{\gamma,\theta,\alpha}(G//K)\hookrightarrow C_{b}(G//K) continuously.

Proof. Let f∈Hγ,θ,αs(G//K)f\in H^{s}_{\gamma,\theta,\alpha}(G//K). Set θx(t,φ)=θ(α(t−1x))\theta_{x}(t,\varphi)=\theta\left(\alpha(t^{-1}x)\right).

(a) For x∈Vx\in V, using ∣φ(x)∣≤1|\varphi(x)|\leq1, the Cauchy-Schwarz inequality and (H), we have

∫S+∫G∣Sθ,αf(t,φ)∣∣θx(t,φ)∣ dt dφ≤∥(1+γ2)s/2Sθ,αf∥L2(∫S+∫G(1+γ(t,φ)2)−s∣θx(t,φ)∣2 dt dφ)12\int_{S^{+}}\int_{G}|S_{\theta,\alpha}f(t,\varphi)||\theta_{x}(t,\varphi)|\,dt\,d\varphi \leq\left\|(1+\gamma^{2})^{s/2}S_{\theta,\alpha}f\right\|_{L^{2}}\left(\int_{S^{+}}\int_{G}(1+\gamma(t,\varphi)^{2})^{-s}|\theta_{x}(t,\varphi)|^{2}\,dt\,d\varphi\right)^{\frac{1}{2}}

(using ∣φ(x)∣≤1|\varphi(x)|\leq1 and the Cauchy-Schwarz inequality)

≤∥f∥Hγ,θ,αs∥h∥L112<∞\leq\|f\|_{H^{s}_{\gamma,\theta,\alpha}}\|h\|_{L^{1}}^{\frac{1}{2}}<\infty

(using (H)).

Let x0∈Vx_{0}\in V and let (xn)(x_{n}) be a sequence which converges to x0x_{0}. Since VV is open, xn∈Vx_{n}\in V for nn large. Put In(t,φ)=Sθ,αf(t,φ)φ(xn)θ(α(t−1xn))I_{n}(t,\varphi)=S_{\theta,\alpha}f(t,\varphi)\varphi(x_{n})\theta\left(\alpha(t^{-1}x_{n})\right). Since there are countably many xnx_{n}, for a.e. (t,φ)(t,\varphi) and all nn, by (H), we have

∣In(t,φ)∣≤∣Sθ,αf(t,φ)∣ ∣θxn(t,φ)∣≤(1+γ2(t,φ))s/2∣Sθ,αf(t,φ)∣h(t,φ)12=:G(t,φ).\begin{aligned} |I_{n}(t,\varphi)|\leq|S_{\theta,\alpha}f(t,\varphi)|\,|\theta_{x_{n}}(t,\varphi)| \\ &\leq(1+\gamma^{2}(t,\varphi))^{s/2}|S_{\theta,\alpha}f(t,\varphi)|h(t,\varphi)^{\frac{1}{2}}=:G(t,\varphi). \end{aligned}

Let us now set

G(t,φ)=(1+γ2(t,φ))s/2∣Sθ,αf(t,φ)∣ ∣h(t,φ)∣12.G(t,\varphi)=(1+\gamma^{2}(t,\varphi))^{s/2}|S_{\theta,\alpha}f(t,\varphi)|\,|h(t,\varphi)|^{\frac{1}{2}}.

We have G∈L1(G×S+)\mathcal{G} \in L^{1}(G \times S^{+}) because (1+γ2)s/2Sθ,αf∈L2(G×S+)(1+\gamma^{2})^{s/2}S_{\theta,\alpha}f \in L^{2}(G \times S^{+}) and h1/2∈L2(G×S+)h^{1/2} \in L^{2}(G \times S^{+}). Moreover, for every (t,φ)(t,\varphi), In(t,φ)→I0(t,φ)I_n(t,\varphi) \to I_0(t,\varphi), because φ\varphi and θ\theta are continuous and so are α\alpha and the group operations. The Lebesgue dominated convergence theorem implies that F(xn)F(x_n) converges to F(x0)F(x_0). Since GG is first countable, FF is continuous at x0x_0.

Now let us show that F=fF=f a.e. on VV. First, FF is KK-bi-invariant on VV: left KK-invariance follows from the change of variables t↦ktt \mapsto kt together with Sθ,αf(kt,φ)=Sθ,αf(t,φ)S_{\theta,\alpha}f(kt,\varphi)=S_{\theta,\alpha}f(t,\varphi) (because ff and φ\varphi are KK-bi-invariant); right KK-invariance follows from φ(xk)=φ(x)\varphi(xk)=\varphi(x), α(K)=K\alpha(K)=K and the right KK-invariance of θ\theta. Now let g∈Cc(G//K)g \in C_c(G//K) with supp⁡g⊂V\operatorname{supp} g \subset V. We have

∫GF(x)g(x)‾ dx=δα12∫G∫S+∫GSθ,αf(t,φ)φ(x)θ(α(t−1x)) dt dφ g(x)‾ dx=δα12∫S+∫GSθ,αf(t,φ)⟨θα,φ,t,g⟩ dt dφ\begin{aligned} \int_G F(x)\overline{g(x)}\,dx=\delta_{\alpha}^{\frac{1}{2}}\int_G\int_{S^+}\int_G S_{\theta,\alpha}f(t,\varphi)\varphi(x)\theta(\alpha(t^{-1}x))\,dt\,d\varphi\,\overline{g(x)}\,dx \\ &=\delta_{\alpha}^{\frac{1}{2}}\int_{S^+}\int_G S_{\theta,\alpha}f(t,\varphi)\langle\theta_{\alpha,\varphi,t},g\rangle\,dt\,d\varphi \end{aligned}

(by Fubini’s theorem)

=⟨f,g⟩.=\langle f,g\rangle.

Hence ∫V(f−F)g‾=0\int_V(f-F)\overline{g}=0 for all such gg. Since f−Ff-F is KK-bi-invariant and locally integrable on VV, f=Ff=F a.e. on VV. This proves (a).

(b) Under (H) everywhere, FF is defined and continuous on GG and equals ff a.e. Under (M), for every x∈Gx \in G, the inequality

∫S+∫G∣Sθ,αf(t,φ)∣∣θx(t,φ)∣ dt dφ≤∥(1+γ2)s/2Sθ,αf∥L2(∫S+∫G(1+γ(t,φ)2)−s∣θx(t,φ)∣2 dt dφ)12\int_{S^+}\int_G |S_{\theta,\alpha}f(t,\varphi)||\theta_x(t,\varphi)|\,dt\,d\varphi\leq\|(1+\gamma^2)^{s/2}S_{\theta,\alpha}f\|_{L^2}\left(\int_{S^+}\int_G(1+\gamma(t,\varphi)^2)^{-s}|\theta_x(t,\varphi)|^2\,dt\,d\varphi\right)^{\frac{1}{2}}

implies

∣F(x)∣≤δα12∫S+∫G∣Sθ,αf(t,φ)∣∣θx(t,φ)∣ dt dφ≤δα12M12∥f∥Hγ,θ,αs.|F(x)|\leq\delta_{\alpha}^{\frac{1}{2}}\int_{S^+}\int_G|S_{\theta,\alpha}f(t,\varphi)||\theta_x(t,\varphi)|\,dt\,d\varphi\leq\delta_{\alpha}^{\frac{1}{2}}M^{\frac{1}{2}}\|f\|_{H^s_{\gamma,\theta,\alpha}}.

Therefore,

∥F∥∞≤δα12M12∥f∥Hγ,θ,αs,\|F\|_{\infty}\leq\delta_{\alpha}^{\frac{1}{2}}M^{\frac{1}{2}}\|f\|_{H^s_{\gamma,\theta,\alpha}},

and the following continuous embedding follows: Hγ,θ,αs(G//K)↪Cb(G//K)H^s_{\gamma,\theta,\alpha}(G//K)\hookrightarrow C_b(G//K).

Example 3.9 (The Euclidean motion group). Let n≥2n \geq2 and (G,K)=(E(n),SO(n))(G,K)=(E(n),SO(n)), where E(n)=Rn⋊SO(n)E(n)=\mathbb{R}^n\rtimes SO(n) with product and inverse

(y,k)(y′,k′)=(y+k⋅y′,kk′),(y,k)−1=(−k−1y,k−1).(y,k)(y',k')=(y+k\cdot y',kk'),\qquad(y,k)^{-1}=(-k^{-1}y,k^{-1}).

The Haar measure of GG is dy dkdy\,dk, where dydy is the Lebesgue measure on Rn\mathbb{R}^n and dkdk is the normalized Haar measure of KK. A function on GG is KK-bi-invariant if and only if it is of the form (y,k)↦f(∣y∣)(y,k)\mapsto f(|y|), i.e. it is a radial function on Rn\mathbb{R}^n [3 Subsection 7.1]. We identify Lp(G//K)L^p(G//K) with the radial functions of Lp(Rn)L^p(\mathbb{R}^n).

The bounded spherical functions are indexed by λ≥0\lambda\ge0:

φλ(y)=Γ(n2)2n−22Jn−22(∣λy∣)(∣λy∣)n−22,∣φλ∣≤1,φλ(0)=1,\varphi_\lambda(y)=\Gamma\left(\frac{n}{2}\right)2^{\frac{n-2}{2}}\frac{J_{\frac{n-2}{2}}(|\lambda y|)}{(|\lambda y|)^{\frac{n-2}{2}}},\qquad|\varphi_\lambda|\le1,\qquad\varphi_\lambda(0)=1,

where JνJ_\nu is the Bessel function of the first kind. They are real-valued, S+≅[0,∞)S^+\cong[0,\infty), and the spherical Fourier transform f^(λ)=∫Rnf(y)φλ(y) dy\widehat{f}(\lambda)=\int_{\mathbb{R}^n}f(y)\varphi_\lambda(y)\,dy is the Hankel transform. With the Haar measure dydy, the Plancherel measure is dφλ=cnλn−1dλd\varphi_\lambda=c_n\lambda^{n-1}d\lambda, where cn>0c_n>0 depends only on nn and on the normalization of the Haar measure.

For a>0a>0 let αa(y,k)=(ay,k)\alpha_a(y,k)=(ay,k). Then αa\alpha_a is a topological automorphism of E(n)E(n) with αa(K)=K\alpha_a(K)=K (it fixes KK pointwise), and d(αa(x))=andxd(\alpha_a(x))=a^n dx, i.e. δαa=an\delta_{\alpha_a}=a^n. Let

θ(y,k)=θ(y)=π−n4e−∣y∣22,\theta(y,k)=\theta(y)=\pi^{-\frac n4}e^{-\frac{|y|^2}{2}},

which is radial, continuous, belongs to L1∩L2L^1\cap L^2, and satisfies ∥θ∥L2=1\|\theta\|_{L^2}=1. For t=(u,h)t=(u,h) and x=(y,k)x=(y,k) in GG,

αa(t−1x)=(ah−1(y−u),h−1k),\alpha_a(t^{-1}x)=\left(ah^{-1}(y-u),h^{-1}k\right),

hence

θ(αa(t−1x))=π−n4e−a2∣y−u∣22,\theta\left(\alpha_a(t^{-1}x)\right)=\pi^{-\frac n4}e^{-\frac{a^2|y-u|^2}{2}},

which depends neither on hh nor on kk. Since φλ\varphi_\lambda and θ\theta are real and the integral over KK is trivial, for a radial ff we get

Sθ,αaf(u,h,λ)=Sθ,αaf(u,λ)=an2π−n4∫Rnf(y)φλ(y)e−a2∣y−u∣22 dy,u∈Rn, λ≥0,(2)S_{\theta,\alpha_a}f(u,h,\lambda)=S_{\theta,\alpha_a}f(u,\lambda)=a^{\frac n2}\pi^{-\frac n4}\int_{\mathbb{R}^n}f(y)\varphi_\lambda(y)e^{-\frac{a^2|y-u|^2}{2}}\,dy,\qquad u\in\mathbb{R}^n,\ \lambda\ge0, \tag*{(2)}

a windowed Hankel-type transform.

Take the weight γ(t,λ)=λ\gamma(t,\lambda)=\lambda, which depends only on the spherical variable. Definition 3.1 gives

Hγ,θ,αas(E(n)//SO(n))={f radial on Rn:cn∫0∞∫Rn(1+λ2)s∣Sθ,αaf(u,λ)∣2 du λn−1dλ<∞}.H^s_{\gamma,\theta,\alpha_a}(E(n)//SO(n))=\left\{f\text{ radial on }\mathbb{R}^n:c_n\int_0^\infty\int_{\mathbb{R}^n}(1+\lambda^2)^s|S_{\theta,\alpha_a}f(u,\lambda)|^2\,du\,\lambda^{n-1}d\lambda<\infty\right\}.

We check the hypotheses of Theorem 3.8. Fix ρ≥0\rho\ge0, r>0r>0 and the KK-bi-invariant open set V={(y,k): ∣∣y∣−ρ∣<r}V=\{(y,k):\ ||y|-\rho|<r\}; every point of GG has such a neighbourhood. For y∈Vy\in V and u∈Rnu\in\mathbb{R}^n we have ∣y−u∣≥∣∣u∣−ρ∣−r|y-u|\ge||u|-\rho|-r whenever ∣∣u∣−ρ∣≥r||u|-\rho|\ge r, hence, with t=(u,h)t=(u,h),

∣θ(αa(t−1y))∣2=π−n2e−a2∣y−u∣2≤π−n2κ(u),κ(u)={1,∣∣u∣−ρ∣<r,e−a2(∣∣u∣−ρ∣−r)2,∣∣u∣−ρ∣≥r.\left|\theta\left(\alpha_a(t^{-1}y)\right)\right|^2=\pi^{-\frac n2}e^{-a^2|y-u|^2}\le\pi^{-\frac n2}\kappa(u),\qquad\kappa(u)= \begin{cases} 1,& ||u|-\rho|<r,\\ e^{-a^2(||u|-\rho|-r)^2},& ||u|-\rho|\ge r. \end{cases}

The function κ\kappa is integrable on Rn\mathbb{R}^n (a bounded annulus and a Gaussian tail). Put h(t,λ)=π−n/2κ(u)(1+λ2)−sh(t,\lambda)=\pi^{-n/2}\kappa(u)(1+\lambda^2)^{-s}: it satisfies (H) on VV, and it is integrable on G×S+G\times S^+ as soon as

∫0∞(1+λ2)−sλn−1dλ<∞.\int_0^\infty(1+\lambda^2)^{-s}\lambda^{n-1}d\lambda<\infty.

Moreover, for every y∈Gy\in G, the change of variables w=t−1yw=t^{-1}y and the unimodularity of GG give ∫G∣θ(αa(t−1y))∣2dt=δαa−1=a−n\int_G|\theta(\alpha_a(t^{-1}y))|^2dt=\delta_{\alpha_a}^{-1}=a^{-n}, so that

M=a−ncn∫0∞(1+λ2)−sλn−1dλ,M=a^{-n}c_n\int_0^\infty(1+\lambda^2)^{-s}\lambda^{n-1}d\lambda,

which is finite if and only if s>n/2s > n/2. Theorem 3.8 therefore shows that, for s>n/2s > n/2, every element of Hγ,θ,αas(E(n)//SO(n))H^{s}_{\gamma,\theta,\alpha_a}(E(n)//SO(n)) has a representative FF which is a bounded continuous radial function on Rn\mathbb{R}^{n}, with

∥F∥∞≤(cn∫0∞(1+λ2)−sλn−1 dλ)12∥f∥Hγ,θ,αas.\lVert F\rVert_{\infty} \leq\left(c_n \int_{0}^{\infty} (1+\lambda^2)^{-s}\lambda^{n-1}\,\mathrm{d}\lambda\right)^{\frac{1}{2}}\lVert f\rVert_{H^{s}_{\gamma,\theta,\alpha_a}} .

The condition s>n/2s > n/2 is the classical threshold for the embedding of Hs(Rn)H^{s}(\mathbb{R}^{n}) into the bounded continuous functions [11] where

Hs(Rn)={u∈S′(Rn):(1+∣ξ∣2)s/2u^(ξ)∈L2(Rn)},H^{s}(\mathbb{R}^{n}) = \left\{u \in\mathcal{S}'(\mathbb{R}^{n}) : (1+\lvert\xi\rvert^{2})^{s/2}\widehat{u}(\xi) \in L^{2}(\mathbb{R}^{n})\right\},

S′(Rn)\mathcal{S}'(\mathbb{R}^{n}) being the space of tempered distributions and u^\widehat{u} is the Fourier transform of uu.

Conclusion

In this paper we have defined Sobolev-type spaces on a Gelfand pair through the Stockwell transform, with the smoothness measured by a weight acting on the time-frequency variables rather than on the frequency alone. The Plancherel-type identity for the Stockwell transform makes these spaces Hilbert spaces, and the usual structural properties of a Sobolev scale hold: continuous embeddings between different smoothness levels and an interpolation inequality. We also obtained embeddings into Lebesgue spaces and into the bounded continuous functions. The embedding into bounded continuous functions rests on a domination condition on the window and requires no equicontinuity of the family of spherical functions. On the Euclidean motion group, the construction leads to a windowed Hankel transform and to the classical smoothness threshold for continuity. Finally, the same construction could be attempted for the Stockwell transform in related frameworks, such as hypergroups or commutative triples, and applied to the regularity of localization operators.

References

Department of Mathematics, University of Lomé, Togo

Email address: [email protected]

Department of Mathematics, University of Kara, Togo

Email address: [email protected]

Department of Mathematics, University of Lomé, Togo

Email address: [email protected]

References

  1. [1]Adams, R.A., Fournier, J.J.F.: Sobolev Spaces (2nd ed.), Vol. 140, Pure and Applied Mathematics, Academic Press, 2003.
  2. [2]Bataka, K.T., Egwe, M. E., Mensah, Y.: Sobolev spaces on hypergroup Gelfand pairs, Far East J. Math. Sci. 142(1), (2025), 57-69.DOI
  3. [3]van Dijk, G.: Introduction to Harmonic Analysis and Generalized Gelfand Pairs, Walter de Gruyter GmbH & Co. KG, Berlin, 2009.
  4. [4]Dosseh, C.G., Todjro, M., Mensah, Y.: The Stockwell transform on Gelfand pairs and localization operators, J. Pseudo-Differ. Oper. Appl. 17(3), (2026), article 55.
  5. [5]Esmaeelzadeh, F.: Localization operators related to Stockwell transforms on locally compact abelian groups, J. Pseudo-Differ. Oper. Appl. 14(4), (2023), article 67.DOI
  6. [6]F. Esmaeelzadeh, The Stockwell transform on locally compact abelian groups, Infinite Dimensional Analysis, Quantum Probability and Related Topics, 2024, vol. 27, no. 1, p.2350002.DOI
  7. [7]Górka, P., Reyes, E. G.: Sobolev spaces on locally compact abelian groups and the bosonic string equation, J. Aust. Math. Soc. 97(1), (2014), 101-114.DOI
  8. [8]Krukowski, M.: Sobolev spaces on Gelfand pairs, arXiv:2003.08519v1.[math.FA], (2020). https://arxiv.org/abs/2003.08519v1.
  9. [9]Kumar, M., Kumar, N.S.: Sobolev spaces on compact groups, Forum Math. 35(4), 2023, 901-911.DOI
  10. [10]Mensah, Y.: Sobolev spaces arising from a generalized spherical Fourier transform, Adv. Math. Sci. 10(7) (2021), 2947-2955.DOI
  11. [11]Stein, E.M., Weiss, G.: Introduction to Fourier Analysis on Euclidean Spaces, Princeton University Press, 1971.
  12. [12]Stockwell, R.G., Mansinha, L., Lowe, R.P.: Localization of the complex spectrum: the S-transform, IEEE Trans. Signal Proc. 44 (4), (1996) 998-1001.DOI
  13. [13]Wolf, J.A.: Harmonic Analysis on Commutative spaces, Amer. Math. Soc., Mathematical Surveys and Monographs 142, 2007.DOI

Paper details

vlnc:W_tem65e9d
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…
Contents