Single annulus \(L^p\) estimates for Hilbert transforms along vector fields (Q373510): Difference between revisions

Let \(v(x)\) be a vector field in \(\mathbb R^2\), and \(H_v\) be the Hilbert transform defined by \[ H_vf(x):=\text{p.v.}\int {f(x-tv(x))}{t^{-1}}dt. \] Then the author and \textit{C. Thiele} showed that if \(v(x_1,x_2)=v(x_1,0)\) and \(v(x_1,0)\) is measurable, then \(\|H_vf\|_p\lesssim \|f\|_p\) for \(p\in(3/2,\infty)\) [Anal. PDE 6, 1577--1600 (2013; Zbl 1285.42014)]. In this paper, the author gives the same estimate for \(p\in(1,\infty)\), under the additional assumption that \(f\) have frequency support in an annulus. His main theorem is as follows: Let \(\tau\) be the trapezoid with corners \((-1/w,1/w)\), \((1/w,1/w)\), \((-2/w,2/w)\) and \((2/w,2/w)\) \((w>0)\), and define \[ \widehat{\Pi_\tau f}(\xi) =\chi_\tau(\xi)\hat f(\xi). \] Let \(v\) be a vector field depending on one variable with slope bounded by \(1\). Let \(p\in(1,\infty)\). Then \(\|(H_v\circ \Pi_\tau\|_p\lesssim \|\Pi_\tau f\|_p\). He gives precise proof of the above theorem, using time-frequency analysis developed by M. Lacey and C. Thiele, and developing some new ideas.