1: \begin{abstract}

2:    Let the  Ornstein-Uhlenbeck process $(X_t)_{t\ge0}$ driven by a fractional Brownian motion $B^{H }$, described    by   $dX_t = -\theta  X_t dt + \sigma dB_t^{H }$ be observed at discrete time instants

3:    $t_k=kh$, $k=0, 1, 2, \cdots, 2n+2 $.  We propose   ergodic type 

4: statistical estimators $\hat \theta_n $, $\hat H_n $ and $\hat \sigma_n $ to estimate all  the parameters $\theta $, $H $ and $\sigma $ in the above 

5: Ornstein-Uhlenbeck  model  simultaneously.  We prove the strong consistence  

6: and the rate of convergence of   the estimators.   The step size $h$

7: can be arbitrarily  fixed and will not be forced to go zero, which is usually a reality.   The tools to use 

8: are the generalized moment  approach (via ergodic theorem) and the Malliavin calculus. 

9: \end{abstract}

10: