作者： R. Rojas-Laguna , O. Ibarra-Manzano , Y. Shmaliy , J. Fernandez-Cortes , A. Marienko
DOI:
关键词: Mathematics 、 Estimator 、 Kalman filter 、 Control theory 、 Moving average 、 Finite impulse response 、 Clock synchronization 、 Mean squared error 、 Linear regression 、 Filter (signal processing)
摘要: Abstract : In this paper, we investigate one of the possibilities to adapt an unbiased moving average (MA) filter (finite impulse response [FIR] filter) slope time error function. The linear regression coefficient is used as a statistical estimator sample slope. We evaluate estimate and present two options for adapting determination. To examine them, generate noisy signal with trend it simple MA, optimally adapted filters. particular errors filters, namely bias, RMSD, RMSE maximal are then compared. Finally mark special features linear-regression- based adaptation.