Adapting product kernels to curves in the plane (Q706098): Difference between revisions

Let \(K\) be a product kernel (that is, a tempered distribution which satisfies differential inequalities and cancellation conditions analogous to those satisfied by the double Hilbert transformation), and let \(p\in (0,\infty)\). Here it is proved that the convolution operator with kernel \(K\) defined on the Schwartz space extends to a bounded operator on \(L^p({\mathbb R}^2)\). { } The main idea of the proof is to decompose \(K\) into a sum of two kernels \(K_i\) (\(i=1,2\)) in such a way that the singularity of \(K_1\) is localized essentially along the coordinate axis \(x=0\), while that of \(K_2\) is localized essentially along a curve \(y=\gamma(x)\). Then it is shown that the operators with kernels \(K_i\) (\(i=1,2\)) are \(L^p\)-bounded by using the curvature properties of \(y=\gamma(x)\), the theory of oscillatory integrals and a suitable modification of an argument of complex interpolation originally developed by \textit{E. M. Stein} and \textit{S. Wainger} [Bull. Am. Math. Soc. 84, 1239--1295 (1978; Zbl 0393.42010)] in order to prove the \(L^p\)-boundedness of the Hilbert transform along curves in the plane.