An Optimal FIR Filtering Algorithm for a Time Error Model of a Precise Clock

Yuriy Shmaliy, Oscar Ibarra-Manzano

Abstract:	This report* is a comparative study of the GPS-based parameter estimates of the precise local clock linear time model. We estimate the time error x and fractional frequency offset y employing the finite impulse response (FIR) digital filters, namely: the simple moving average (MA), the second order low pass (LP), and the optimally unbiased (OU). The estimates are provided analytically, and we show that, in terms of a minimum root-mean-square error (RMSE), the simple MA is best to use for x estimates when 0 1 = y < y , the second-order LP FIR is best if 1 2 y < y < y , and the OU is best for y < y 2 , where 1 y and 2 y are determined constants. Finally, we present an optimal FIR algorithm for the precise clock linear time error model and show that of all the estimators, including the Kalman, the presented algorithm yields minimum RMSE of the fractional frequency offset
Published in:	Proceedings of the 34th Annual Precise Time and Time Interval Systems and Applications Meeting December 3 - 5, 2002 Hyatt Regency Reston Town Center Reston, Virginia
Pages:	527 - 538
Cite this article:	Shmaliy, Yuriy, Ibarra-Manzano, Oscar, "An Optimal FIR Filtering Algorithm for a Time Error Model of a Precise Clock," Proceedings of the 34th Annual Precise Time and Time Interval Systems and Applications Meeting, Reston, Virginia, December 2002, pp. 527-538.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In