\(\ell^p(\mathbb{Z}^d)\)-estimates for discrete operators of Radon type: maximal functions and vector-valued estimates (Q2317976)

The authors study discrete analogues of operators of Radon type. The approach undertaken unifies (and extends to vector-valued versions) results of two major streams of the theory, the Bourgain maximal theorems [\textit{J. Bourgain}, Publ. Math., Inst. Hautes Étud. Sci. 69, 5--45 (1989; Zbl 0705.28008)], and the theorem of Ionescu-Wainger for singular Radon transforms [\textit{A. D. Ionescu} and \textit{S. Wainger}, J. Amer. Math. Soc. 19 (2), 357--383 (2006; Zbl 1158.42007)]. The main results of this paper can be stated as follows: Let \(K \in\mathcal{C}^1(\mathbb R^k \setminus\{0\})\) be a suitable Calderón-Zygmund kernel and let \(\mathcal P=(\mathcal P_1,\dots,\mathcal P_{d_0}): \mathbb Z^k\rightarrow \mathbb Z^{d_0}\) be a polynomial mapping, where for each \(j \in \{1,\dots,d_0\}\) the function \(\mathcal P_j: \mathbb Z^k \rightarrow\mathbb Z\) is a polynomial of \(k\) variables with integer coefficients such that \(\mathcal P_j(0)=0\). Let \[ \mathcal{T}^{\mathcal P}_Nf(x)=\sum_{y\in \mathbb Z^k_N\setminus\{0\}}f(x-\mathcal P(y))K(y) \] be the discrete truncated singular Radon transform, where \(\mathbb Z^k_N=\{-N,\dots,-1,0,1,\dots,N\}^k\). Let \[ {M}^{\mathcal P}_Nf(x)=N^{-k}\sum_{y\in \mathbb N^k_N}f(x-\mathcal P(y)) \] be the Radon averaging operator, where \(\mathbb N^k_N=\{1,2,\dots,N\}^k\). Theorem C. For every \(p\in(1,\infty)\) there is \(C_p >0\) such that for all \((f_t :t\in\mathbb N) \in \ell^p(\ell^2(\mathbb Z^{d_0}))\) we have \[ \bigg\|\bigg(\sum_{t\in\mathbb N}\sup_{N\in\mathbb N}|M_N^{\mathcal P}f_t|^2\bigg)^{1/2}\bigg\|_{\ell^p}\le C_p \bigg\|\bigg(\sum_{t\in\mathbb N}|f_t|^2\bigg)^{1/2}\bigg\|_{\ell^p}. \] Moreover, the constant \(C_p\) is independent of the coefficients of the polynomial mapping \(\mathcal P\). Theorem D. For every \(p\in(1,\infty)\) there is \(C_p >0\) such that for all \((f_t :t\in\mathbb N) \in \ell^p(\ell^2(\mathbb Z^{d_0}))\) we have \[ \bigg\|\bigg(\sum_{t\in\mathbb N}\sup_{N\in\mathbb N}|T_N^{\mathcal P}f_t|^2\bigg)^{1/2}\bigg\|_{\ell^p}\le C_p \bigg\|\bigg(\sum_{t\in\mathbb N}|f_t|^2\bigg)^{1/2}\bigg\|_{\ell^p}. \] Moreover, the constant \(C_p\) is independent of the coefficients of the polynomial mapping \(\mathcal P\). The proof of theorems C and D is based on an analysis of the Fourier multipliers corresponding to the operators \(M_{2^n}^{\mathcal P}\) and \(T_{2^n}^{\mathcal P}\), respectively. There is a final Appendix devoted to provide some vector-valued estimates for maximal operators of Radon type in the continuous settings.