Identification of first-order plus dead-time model from less step response data

This study presents an improved method for first-order plus dead-time (FOPDT) model identification from less step response data. Firstly, B-spline series expansions are used to approximate step responses, providing effective interpolation values for modeling computation. Then, to enhance modeling accuracy, a least-squares method with a regressive compensation scheme is proposed, which adaptively adjusts error weight coefficients to minimize the response deviation between identified model and actual process. The proposed identification method is not only suitable for cases of less sampling data, but also for those of non-uniform sampling data (where most existing methods of transfer function model identification cannot be applied directly). Simulation results illustrate the effectiveness of the proposed approach.