Robust Gradient Iterative Estimation Algorithm for ExpARX Models With Random Missing Outputs

This study presents a LookAhead-RAdam gradient iterative algorithm to identify ExpARX models with random missing outputs. The LookAhead-RAdam gradient iterative algorithm is used to optimize the step size of each element and adjust the direction to effectively update the ExpARX model parameter estimation through the estimated outputs. Compared to the classical gradient iterative algorithm, this study improves the estimation accuracy of the missing outputs and the parameter estimation convergence rate by introducing the LookAhead algorithm and RAdam algorithm. To validate the algorithm developed, a series of bench tests were conducted with computational experiments. Finally, the effectiveness of the proposed design approach is demonstrated by a simulation example.