Iterative constrained penalized likelihood estimation of parameters for CDMA
Ejaz Khan, Dirk T. M. Slock · 2003
We describe an iterative method for maximum likelihood (ML) parameter estimation corrupted by additive white Gaussian noise. In the objective function we subtract/add the Kullback-Leibler (KL) distance function or Euclidean distance function to keep the old parameter set close to the new ones and it can be considered as a penalty term. The above augmented cost function can be maximized/minimized over the constraint that the detected data vector lie on the sphere. We simplify this constraint function by using first order Taylor expansion at the old parameter value. The useful behavior of the proposed algorithm is verified by numerical experiments.