Robust Active Disturbance Rejection Control Based on Optimized Extended State Observer via Adaptive Neural-Network for Induction Motor Drive

In this paper, a robust adaptive control scheme using feedback linearization control (FLC) technique and active disturbance rejection control (ADRC) to achieve strong anti- disturbance performance for induction motor (IM) drive. Firstly, a FLC is developed to ensure the IM drive stability. As a result of model uncertainties, unmodeled dynamics and exogenous disorder, the control system performance can be affected. Aiming to improve the anti-disturbance performance, an extended state observer (ESO) is developed for estimating the total disturbances and states of the IM drive. To optimize the behavior of the ADRC, an action dependent heuristic dynamic programming (ADHDP) online strategy is employed for self- tuning of ESO parameters optimally and iteratively improving the estimation accuracy. Thus, critic and actor neural networks are adopted to solve the Hamilton-Jacobi-Bellman (HJB) equation online. At the same time, actor neural network provides the optimal control performance. The validity of the proposed FLC-ADRC scheme based on the optimal adaptive ESO via ADHDP approach is verified experimentally. The test results verify that the FLC-ADRC can make the IM drive follow the reference trajectory quickly and accurately and grants robust control performance regardless of uncertain dynamics as well as interior and exterior disorders.