ANN-Based Control Strategy for Improving Battery Charger Dynamics in Personal Mobility Devices.

This study proposes the use of an Artificial Neural Network (ANN)-based control strategy to improve the dynamic performance of battery chargers used in Personal Mobility Devices (PMDs), such as electric scooters, bicycles, and skateboards. Traditionally, Proportional-Integral (PI) controllers are used to regulate the output voltage of battery chargers. However, PI controllers require careful gain tuning, and higher gains—while improving response time—can lead to overshoot and instability during transient conditions. To address these limitations, this work introduces an ANN-based voltage regulation method for PMD battery chargers. Unlike conventional controllers, ANN models can learn from system behavior and adapt in real time, offering superior control in nonlinear and time-varying environments. By leveraging past data and dynamic feedback, the ANN controller anticipates voltage deviations and adjusts control actions accordingly.Simulation results validate the effectiveness of the proposed ANN approach, showing improved voltage regulation, faster settling time, and reduced overshoot under dynamic operating conditions. These outcomes confirm the ANN controller’s potential as a high-performance alternative to traditional PI-based methods in PMD battery charger applications.