On the lacunary spherical maximal function on the Heisenberg group (Q2215834)

Let \(\mathbb H^n\) be the Heisenberg group whose underlying manifold is \(\mathbb C^n \times \mathbb R\) with the group law \[(z, t) (w, s) = \Bigl(z+w, t+s+\frac{1}{2} \Im(z\bar{w})\Bigr).\] The group \(\mathbb H^n\) admits a family of dilations \(\{\delta_r: r> 0\}\) acting on \(\mathbb H^n\) as automorphisms, defined by \[\delta_r(z, t) = (rz, r^2t).\] The Koranyi norm on \(\mathbb H^n\), defined by \(|(z, t)| = (|z|^4 + 16t^2)^{1/4}\), satisfies \(|\delta_r(z, t)| = r|(z,t)|.\) Let \(S_K\) be the unit sphere with respect to this norm: \(S_K = \{(z, t):~|(z, t)| = 1\}.\) Then, there is a unique Radon measure \(\sigma\) on \(S_K\) such that \[\int_{\mathbb H^n} f(z, t)\,dz\,dt = \int_0^\infty\!\int_{S_K} f(\delta_r w) \,d\sigma(w)~r^{Q-1}\,dr\] where \(Q = 2n+2\) is the homogenous dimension. Let \(\delta_r \sigma = \sigma_r.\) Then the spherical mean value operator is defined by \[A_rf(x) = f \ast \sigma_r(x) = \int_{|y|=1}f(x \cdot \delta_ry^{-1})~dy\] and the associated spherical maximal operator is \[M_{\mathbb H^n}f(x) = \sup_{r>0}|A_rf(x)|.\] A well-known result due to Cowling states that the operator \(M_{\mathbb H^n}\) is bounded on \(L^p(\mathbb H^n)\) for all \(p > \frac{2n+1}{2n}.\) The paper under review proves a lacunary version of the above result. This is an analogue of C. P. Calderon's theorem for the lacunary spherical means on the Euclidean spaces (which says that the lacunary spherical maximal function is bounded on \(L^p(\mathbb R^n)\) for all \(1 < p < \infty\)). Fix a \(\delta \in (0, \frac{1}{96}),\) and define the lacunary spherical maximal function on \(\mathbb H^n\) by \[M_{\mathbb H^n}^{\mathrm{lac}}f(x) = \sup_{k \in \mathbb Z} |A_{\delta^k}f(x)|.\] Theorem: Let \(n \geq 2.\) The lacunary spherical maximal function \(M_{\mathbb H^n}^{lac}\) is bounded on \(L^p(\mathbb H^n)\) for all \(1 < p < \infty.\) The key step is a sparse domination bound for the lacunary spherical maximal function: Let \(1 <p, q < \infty\) be such that \((\frac{1}{p}, \frac{1}{q})\) belong to the interior of the triangle joining \((1, 0), (0, 1)\) and \((\frac{2n}{2n+1}, \frac{2n}{2n+1}).\) Then for any pair of compactly supported bounded functions \((f, g)\) there exists a \((p, q)\)-sparse form \(\Lambda_{S, p, q}\) such that \[\langle M_{\mathbb H^n}^{\mathrm{lac}}f, g \rangle \leq C\Lambda_{S, p, q}(f, g).\] Key ingredients in proving the above sparse bounds are (i) the \(L^p\) improving property of \(A_r\), (ii) the continuity property of \(A_r.\) If \( (\frac{1}{p}, \frac{1}{q})\) is as above or lies on the straight line joining the points \((0, 0)\) and \((1,1),\) the \(L^p\)-improving property states that \[\|A_rf\|_q \leq C r^{Q(\frac{1}{q}-\frac{1}{p})} \|f\|_p.\] The continuity property gives the estimate \[\|A_r - A_r \tau_a\|_{L^p \to L^q} \leq C r^{Q(\frac{1}{q}- \frac{1}{p})} |\delta_r^{-1}a|^\eta\] for some \(0 < \eta < 1\), where \(\tau_a\) stands for the left translation by \(a\in \mathbb H^n.\) Using the sparse bounds and well-known boundedness properties of sparse forms the main result is established.