Enhancing Reference Tracking in Induction Motor Drives: A Hybrid Model Predictive Control and Artificial Neural Network Approach

ABSTRACT Induction motors (IMs) are widely used in industry for their robustness and cost‐effectiveness, yet their nonlinear and multivariable dynamics pose significant challenges for high‐performance speed control. Model Predictive Control (MPC) is a well‐established solution but suffers from limitations related to model accuracy and parameter uncertainty. Artificial Neural Networks (ANNs) offer an alternative by capturing complex nonlinear behaviors without extensive parameterization. This paper presents and experimentally validates a hybrid control strategy that integrates MPC with an ANN‐based model for IM drives. The proposed ANN‐MPC employs a NARX neural network to enhance prediction accuracy and improve control performance. A key contribution of this work is the comprehensive evaluation of the controller on a real‐time platform, including comparisons with conventional MPC and PI controllers under diverse operating conditions. Experimental results demonstrate superior trajectory tracking, reduced overshoot, faster settling times, and improved robustness to parameter variations and disturbances, while maintaining real‐time computational feasibility. Additionally, the manuscript includes a dedicated discussion on stability and robustness considerations supported by experimental evidence, highlighting the practical viability of ANN‐based predictive control for nonlinear systems.