Least square method for complex estimation
Bowen Cui · Journal of Anhui University · 2005
The paper presents a new U-D factorization based least squares methods for complex estimation. In the conventional weighted recursive least squares algorithm with forgot factors for complex-domain data, the covariance matrix P(k) converges to zero as time index k tends toward infinity. In order to keep converge in parameters estimation, the covariance matrix P(k) is to be decomposed using U-D factorization of matrix, the recursive calculation of P(k) is turned to be recursive calculation of U(k) and D(k),and the stability of numerical calculation is maintained.