This paper addresses the time-delay estimation problem of SISO systems using uniformly sampled input-output data. A dependence measure known as Schweizer and Wolff's σ (SWσ) from the copula theory is employed to measure the dependence of the input-output data. The quantitative relation between the time-delay and SWσ is studied, and the time-delay is estimated given that the SWσ of the input-output data reaches its maximum value when the time-delay is removed. The major advantage of our work is that no other parameters of the system must be estimated before or during the time-delay estimation process. Experiments and comparisons are used to verify the feasibility and effectiveness of the proposed method.
Keywords: Copula; Dependence measure; SWσ; Time-delay estimation.
Copyright © 2018 ISA. Published by Elsevier Ltd. All rights reserved.