On mean curvature functions of Brownian paths (Q860704)

Consider a standard Brownian motion \((W_{t})_{t\geq 0}\) in \(\mathbb{R}^{d}\), \(d\geq 2\), and let \(Z^{t}=\{ W_{s}:0\leq s\leq t\}\), \(t\geq0,\) which is a random closed set. This allows to consider the (random) support (curvature) measures \(\mu _{0}( Z^{t};\cdot) \) and \(\mu _{1}( Z^{t};\cdot) \) as defined by \textit{D. Hug}, the author and \textit{W. Weil} [Math. Z. 246, No. 1/2, 237--272 (2004; Zbl 1059.53061)]. These are signed measures on \(\mathbb{R}^{d}\times S^{d-1},\) where \(S^{d-1}\) is the unit sphere in \(\mathbb{R}^{d}\), that are supported by the normal bundle of \( Z^{t} \). Next, define the local reach \(\delta ( Z^{t},x,u) \) at a point \(( x,u) \in Z^{t}\times S^{d-1}\) as the largest \(s>0\) such that an open ball of radius \(s\) centered at \(x+su\) does not intersect \( Z^{t}\). The author studies two mean curvature functions of Brownian motion defined as \(F_{i}( t,s) =E[ \int \mathbf{1}\{ \delta (Z^{t};x,u) >s\} \mu _{i}( Z^{t};d(x,))]\), \(i=0,1\). An equivalent definition is available in terms of the Wiener sausage.