A Novel Iterative System Identification and Modeling Scheme With Simultaneous Time-Delay and Rational Parameter Estimation

Abstract

In this work, a new method is proposed to identify systems in the form of first and second-order continuous time-delayed (CTD) models where time-delay and rational parameters are estimated simultaneously. The time delay is explicitly brought into system parameters by discretizing the CTD models. The sampled input-output data from the system is used in an iterative prediction error minimization (PEM) algorithm to identify the discrete-time delayed (DTD) model from which the equivalent CTD model is extracted. The DTD model includes the fractional and integer part of time-delay, where the phase contribution due to the fractional part is used to correct the estimation of delay as well as the termination of the algorithm until the phase contribution lies within one sample time. The efficacy of the proposed method is demonstrated by the simulation study with noisy measurements for four different systems along with the experimental validation.

Keywords

• System identification
• time-delay estimation
• prediction error methods
• iterative algorithms
• continuous time-delayed systems
• discrete time-delayed systems