Blind-deconvolution analysis of transient curves by the use of a convolved autoregressive model

Abstract

A new method is proposed for analyzing a transient curve as a multiexponential decay convolved with an instrumental response function and for estimating the decay constants and amplitudes without knowledge of the response function. The method requires a single-exponential reference curve observed with the same response function as that for a sample curve, whereas its decay-constant and amplitude values are not necessary for the estimation. The sample and reference curves are analyzed with a convolved autoregressive model, which is based on a linear least-squares algorithm. This method has the advantage of a short computation time when compared with a conventional nonlinear least-squares fitting technique. Double-exponential curves are also acceptable as the reference, if two curves with different decay parameters can be provided for the analysis.