Another approach to Brownian motion (Q2490061)

Let \(f:\mathbb{R}\to\mathbb{R}\) be a symmetric, continuous, convex function, with \(f(0)=0\), strictly increasing on \(\mathbb{R}^+\), satisfying \(f(Kx)\leq K^p f(x)\), for some \(p\in [1,2)\) and \(K\geq 0\). The main result of the present paper is that if some process \((X(t),t\in[0,1])\) with independent increments satisfies \(X(0)=0\), \(Ef (X(1))<\infty\) and, \(\mu\)-a.s. for all \(s\in[0,1]\), \[ \liminf_{h\searrow 0}Ef (h^{-1/2}[X(s+h)-X(s) ])\geq Ef(X(1)), \] then \(X (t)\) is a Gaussian process that admits the representation \[ X(t)=\sigma W(t) +E X(t),\quad t\in [0,1], \] where \(W(t)\) is a Brownian motion and \(\sigma\) a constant. The authors derive a characterisation of Brownian motion. Further an application to the central limit theorem is given.