Minimum variance filtering with the linearly constrained inverse QRD-RLS algorithm

A general linearly constrained recursive least squares (RLS) filtering algorithm, based on an inverse QR decomposition, is developed and applied to the minimum variance filtering problem, where the adaptation (or Kalman) gain is evaluated via the Givens rotation. Also, the LS weight vector can be computed without back substitution and achieve fast convergence and good numerical properties. The numerical stability of the proposed method, in terms of constrained drift is emphasized. We show that it outperforms the method using the fast linearly constrained RLS algorithm and its modified version