XGBoost-Based Intelligent Multi-Source Coordination in an Electric Vehicle Employing a Super-Boost Power Converter

Abstract

The central challenge related to the development of electric vehicles (EVs) involves the effective integration of multiple input sources to create a robust and efficient power system. Recent advancements have focused on optimizing the power distribution within hybrid systems that combine batteries, supercapacitors, and renewable sources like solar PV. Conventionally, energy management systems (EMSs) have relied on rule-based algorithms or deterministic optimization methods. However, these techniques often lack adaptability under real-world driving conditions and face significant challenges regarding their generalizability and computational complexity when applied to dynamic driving cycles. Machine learning approaches are capable of modeling the complex, non-linear interactions between multiple energy sources to ensure intelligent power coordination. This paper proposes a novel Extreme Gradient Boosting (XGBoost)-based intelligent EMS for a BLDC motor-driven electric vehicle (e-bike) utilizing a hybrid battery–solar configuration with a supercapacitor for regenerative braking. The proposed system integrates a super-boost converter for efficient multi-source power delivery. The results show accurate energy source identification, effective multi-source coordination, improved energy utilization, reduced battery stress, and a rapid decision-making capability, which prove the feasibility of the proposed method for real-time electric bicycle energy management. The simulation was executed using the MATLAB/Simulink platform, and the obtained results are outlined.