Time-delay estimation of nonlinear systems with delay in states and outputs

Abstract

Aim of this paper is to estimate time delays in nonlinear dynamic systems with unknown time-delays, such that time-delays appear in both state and output equations. The problem of estimating unknown delays is formulated as a dynamic optimization problem in which the cost function is defined based on the least-squares error between predicted and actual system output. Using proposed estimation method derives unknown time-delays by evaluating the gradient of cost function. Since, the gradient cannot be obtained analytically by standard differentiation rules, a practical computational method is presented to compute gradient and related cost function. The effectiveness of proposed algorithm is finally demonstrated by applying it to a practical example. © 2016 IEEE.