Conditioned limit theorems for products of positive random matrices

Inspired by a recent paper of I. Grama, E. Le Page and M. Peigne (Grama et al., 2014), we consider a sequence (g(n))(n >= 1) of i.i.d. random d x d-matrices with non-negative entries and study the fluctuations of the process (log vertical bar g(n) ... g(1)x vertical bar)(n >= 1) for any non-zero vector x in R-d with non-negative coordinates. Our method involves approximating this process by a martingale and studying harmonic functions for its restriction to the upper half line. Under certain conditions, the probability for this process to stay in the upper half real line up to time n decreases as c/root n for some positive constant c.