Identification of Output Error Model for Linear Time Delay Systems Subject to Load Disturbance with Unknown Dynamics

The paper proposes a discrete-time output error model (OEM) identification method for linear time delay systems subject to load disturbance with unknown dynamics. By viewing the load disturbance response as a dynamic parameter to be identified, two recursive least squares (RLS) algorithms are established for simultaneous estimation of model parameters and tracking of load disturbance characteristics. Meanwhile, another unknown parameter namely time delay is derived using a one-dimensional searching strategy. An auxiliary model is utilized for recursive estimation of the noise-free system output against stochastic noise. Furthermore, two adaptive forgetting factors are introduced to improve the convergence rate of parameter estimation and the tracking performance of load disturbance response. An illustrative example is shown to demonstrate the effectiveness of the proposed identification method.