MULTIGHOST CANCELLATION TECHNIQUE FOR HDTV SYSTEMS, USING A DERIVATION OF THE OLS LEARNING ALGORITHM

In this paper, very successful multiple video ghost cancellation simul ations have been obtained with the application of a designed learning algorithm, based on the Orthogonal Least Square method [1,2], to a cha nnel identification process based on a FIR system model The algorithm can be visualized by assuming the existence of a data matrix, at the i nput of the equalizer, which is created with a time shifting process, in order to generate ''M'' column vectors. The designed a algorithm pr ocesses these vectors and, with the aid of an orthogonalization method , calculates a set of the most representative one, with respect to a d esired output signal. The delays and amplitudes of the ghosts were obt ained;vith the aid of a forward regressor method. The process has show n to be very effective for the accurate calculation of the ghost param eters, even in the presence of considerable noise levels, and is also used to train REP approximation networks for the systematic selection of its centroids [2]. In all the tests performed in this work, the pro posed technique has given much better results than using conventional algorithms [3,4,7].