Improving threshold performance of the IQML algorithm
W.J.L. Read · IEEE Transactions on Signal Processing · 2000
In this article, improvements to the iterative quadratic maximum likelihood (IQML) algorithm are presented. The modifications are twofold: modification of the nontriviality constraint on the polynomial coefficient vector and the addition of a noise preprocessing step based on the eigendecomposition of the data covariance matrix. Simulation results are provided.