Yun-Da HuangSOO-CHANG PEIJong-Jy Shyu2018-09-102018-09-102009-0315497747http://scholars.lib.ntu.edu.tw/handle/123456789/351872https://www.scopus.com/inward/record.uri?eid=2-s2.0-64049118760&doi=10.1109%2fTCSII.2008.2011598&partnerID=40&md5=326080d29e6a6ee36891694cde8eab15In this brief, a new coefficient relationship is proposed for the design of variable fractional-delay (VFD) finite-impulse response (FIR) filters by the weighted least-squares (WLS) method so that the number of filter coefficients to be designed can approximately be halved. To reduce the computational cost, closed-form expressions for the elements of related vectors and matrices are derived. Several design examples are presented, and the comparisons show that the overall performance of the proposed method is comparable with that of the conventional method; however, the number of filter coefficients to be designed for the proposed method is about half of that in the conventional method. © 2009 IEEE.Variable filters; Variable fractional-delay (VFD) finite-impulse response (FIR) filter; Weighted least-squares (WLS) approachImpulse response; Least squares approximations; Closed-form expression; Computational costs; Conventional methods; Filter coefficients; Variable filters; Variable fractional delay; Weighted least squares; Weighted least squares approach; FIR filtersjournal article10.1109/tcsii.2008.20115982-s2.0-64049118760WOS:000264628900008