Central limit theorems for radial random walks on matrices for

Abstract. Let be a fixed probability measure. For each dimension , let be i.i.d. -valued radial random variables with radial distribution . We derive two central limit theorems (CLTs) for for with normal limits. The first CLT for follows from known estimates of convergence in the CLT on , while the second CLT for will be a consequence of asymptotic properties of Bessel convolutions. Both limit theorems are considered also for -invariant random walks on the space of matrices instead of for and fixed dimension .