Online Optimization of the Laguerre Filters Using Backtracking Search Algorithm with Filter Weight Adjustment

In this paper, Backtracking Search Algorithm (BSA) is combined with a Filter Weight Adjustment (FWA) method to optimize free parameters of Laguerre filters. The proposed method alters the filter weights and is capable of extracting pole locations by minimizing the Mean-Squared Error (MSE) between impulse response of a given discrete system and its Laguerre approximation. Simulation results show that BSA with FWA gives roughly 20dB and 10dB reduction in MSE compared to gradient-based parameter estimation methods and BSA, respectively.

کلیدواژه ها:

Laguerre Filter ، Filter Weights ، Evolutionary Algorithm ، Backtracking Search Algorithm ، Gradient-based methods ، Mean Squared Error ،