Tuning a Generalized Model Predictive Controller for a Boost Converter in a PV System Employing a Genetic Algorithm in an IoT Environment

This paper proposes an intelligent tuning framework for a boost converter connected to a photovoltaic (PV) system using a generalized Model Predictive Control (MPC) strategy integrated with an Internet of Things (IoT) architecture. The core contribution lies in the real-time optimization of the MPC weighting matrices through a Genetic Algorithm (GA) executed on a remote server, while the converter operates on a physical experimental platform. The proposed IoT-based framework enables bidirectional data exchange, where real-time voltage and current measurements from the solar plant are used to evaluate the GA fitness function, and the optimized control parameters are transmitted back to the controller. This approach mitigates modeling inaccuracies and parameter uncertainties commonly associated with offline and simulation-based tuning methods. Experimental results obtained using a solar development kit demonstrate that the proposed method significantly improves dynamic performance, achieving reduced overshoot and settling time when compared with deterministic MPC tuning and conventional control strategies, including SMC, SM+PI, and 2P2Z controllers. The results confirm the effectiveness, robustness, and practical feasibility of combining predictive control, evolutionary optimization, and IoT connectivity for high-performance control of PV power converters.