AbstractA fiberoptic strain sensor is being developed that exploits the fact that the difference between the central frequency of Brillouin scattering, which is a type of back scattering that occurs in optical fibers, and the incident light frequency depends on the strain of the optical fiber. The measurement of the central frequency of the Brillouin scattering requires an estimation of the power spectrum distribution of the Brillouin scattering. An improvement in distribution estimation accuracy can potentially improve both the accuracy of central frequency measurements and the precision of strain measurements. Focusing on the problem of the power spectrum distribution of the Brillouin scattering, this paper proposes a weighted repetition least mean square algorithm. The paper also examines the stability of the proposed repetition algorithm and demonstrates the effectiveness of the proposed approach through simulations. © 2003 Wiley Periodicals, Inc. Electron Comm Jpn Pt 1, 87(1): 72–80, 2004; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecja.1183