Weighted least squares estimator for the squared radial Ornstein-Uhlenbeck process

We estimate simultaneously both dimensional and drift parameters of a squared radial Ornstein-Uhlenbeck process. We do not restrict ourself to the case where the process never reaches zero. In order to avoid the use of unmanageable stopping times and natural but intractable estimator, we propose to make use of a weighted least squares estimator. We establish strong consistency and asymptotic normality for this estimator.