Transversal families of nonlinear projections and generalizations of Favard length (Q6046073): Difference between revisions

The Favard length of a set \(E\subset\mathbb{R}^2\) is defined by \(\mathrm{Fav}(E)=\frac{1}{\pi}\int_0^{\pi}|P_{\theta}(E)|d\theta\), where \(P_{\theta}\) is the orthogonal projection into a line \(L_{\theta}\) through the origin at angle \(\theta\) from the positive \(x\)-axis and \(|\cdot|\) denotes the 1-dimensional Hausdorff measure. The aim of the paper is to prove a generalization of a theorem due to \textit{P. Mattila} [Indiana Univ. Math. J. 39, No. 1, 185--198 (1990; Zbl 0682.28003)] concerning the Favard length for neighbourhoods in the case when families of projections are not orthogonal projections. It is assumed that the families of projections satisfy a so-called s-transversality condition (def. 1.2 and 1.3), which means some compatibility condition for different parameters. The most important case considered in the paper is when \(s=m\), where \(m\) is the dimension of the target space. The natural example of a transversal family is the collection of orthogonal projections from \(\mathbb{R}^n\) to \(\mathbb{R}^m\) for some \(m<n\). Another example is the family of radial maps defined in the following way: for a point \(a\in\mathbb{R}^n\) the radial projection with a center \(a\) maps \(\mathbb{R}^n\diagdown\{a\}\) into \(S^{n-1}\) via \(P_a(x)=\frac{x-a}{|x-a|}\). The generalization of Mattila's theorem is the following result Theorem 1.5. [average nonlinear projection length for neighborhoods] Assume that \(\{ \widetilde{\pi_{\alpha}} : \alpha \in A \}\) is an \(m\)-transversal family of projections into an \(m\)-dimensional space. Fix a positive Borel probability measure \(\mu\) supported on a compact set \(F\subseteq\Omega\), so that \[ \mu(B(x,r))\lesssim r^t \] for all \(x\in\Omega\) and \(0<r<\infty\). \begin{itemize} \item If \(t<m\), then \[ \int_A h(\widetilde{\pi_{\alpha}}F(r))d\psi(\alpha)\gtrsim r^{m-1}. \] \item If \(t=m\), then \[ \int_A h(\widetilde{\pi_{\alpha}}F(r))d\psi(\alpha)\gtrsim (\log r^{-1})^{-1}. \] \end{itemize} The authors have introduced also the notion of the Favard surface length for sets is higher dimensions and proved the following theorem (for dimension 3 with the observation that the change for greater dimension offers no special difficulty): Theorem 1.10. [Favard surface length of neighborhoods.] Let \(E\) be a compact set in the plane and \(\Gamma\) denote a surface in \(\mathbb{R}^3\) defined by \(\Gamma=\{ (t,\gamma(t)) : t\in I \}\), where \(\gamma : \mathbb{R}^2 \to \mathbb{R}\), \(\gamma(s) = f(|s|)\), and \(f : \mathbb{R} \to \mathbb{R}\) is a \(C^2\) function on a nonempty compact interval \(I\) satisfying \(f(x) = f(-x)\), with \(f''>0\) on \(I\). Assume further that \(E\) supports a Borel probability measure \(\mu\) with the \(t\)-dimensional growth condition \(\mu (B(x,r))\lesssim r^t\) for all \(x \in E\), \(0 < r < \infty\). The following statements hold: \begin{itemize} \item The family of curve projections \(\Phi_{\alpha}\) is 1-transversal. \item If \(t < 1\), then for all sufficiently small \(r\) we have \[ \mathrm{Fav}_{\Gamma}(E(r))\gtrsim r^{1-t}. \] \item If \(t = 1\), then for all sufficiently small \(r\) we have \[ \mathrm{Fav}_{\Gamma}(E(r))\gtrsim (\log r^{-1})^{-1}. \] \end{itemize} The notation \(A\lesssim B\) means always that there exists a constant \(C\) so that \(A \leq C \cdot B\) and \(A \sim B\) means that \(A \lesssim B\) and \(B \lesssim A\).