Implementation of a Battery Management System for Electric Vehicles Based on Hybrid Discriminative RBM Approach

Abstract

All sorts of high-power devices rely on batteries; electric vehicles and hybrid electric vehicles are only two examples. The secure and dependable functioning of the batteries depends on a good BMS. Charging, state estimation, and battery modeling are some of the technical parts of BMSs that are briefly touched upon in this proposed. Choosing a method, doing preprocessing, extracting features, and training the model are all steps that must be executed in precisely the correct order. It comprises relative SOC and based SOC, and it may be easily calculated in preprocessing using the relationship between the energy level and the battery's properties. Principal component analysis (PCA) is a mathematical technique used in feature selection that determines a number of linked variables that account for as much variability in the data as feasible by reducing the number of uncorrelated variables. Using Hybrid Discriminative RBM, we trained the model. An impressive 96.88% accuracy was achieved, according to the results.