We consider a linear filtering model (with feedback) when the
  observation noise is an Ornstein-Ulhenbeck process with parameter
  $\beta$. The coefficients appearing in the model are all assumed to
  be bounded. In addition, the coefficients appearing in the
  observation equation are also assumed to be differentiable. We
  consider the genral case when the OU noise is also correlated with
  the signal. Under these conditions we derive the filtering equations
  for the optimal filter.