A Fast Lane Approach to LMS prediction of respiratory motion signals

As a tool for predicting stationary signals, the Least Mean Squares (LMS) algorithm is widely used improvement, the family of normalised LMS algorithms, is known to outperform this algorit However, they still remain sensitive to selecting wrong parameters, being the learning coefficientm the signal history length M. We propose an improved version of both algorithms using a Fast L Approach, based on parallel evaluation of several competing predictors. Thesewere applied to respira motion data from motion-compensated radiosurgery. Prediction was performed using arbitr selected values for the learning coefficientm2 0;0:3 and the signal history lengthM2 1⁄21;15 . The res were compared to prediction using the globally optimal values of m and M found using a grid sea When the learning algorithm is seeded using locally optimal values (found using a grid search on the 96 s of data), more than 44% of the test cases outperform the globally optimal result. In about 38% of cases, the result comes to within 5% and, in about 9% of the cases, to within 5–10% of the global optim This indicates that the Fast Lane Approach is a robust method for selecting the parameters m and 2008 Elsevier Ltd. All rights reser