Data-Driven Energy Consumption Forecasting In Electric Bus Systems Using Machine Learning

Abstract

This paper examines the utilisation of machine learning techniques to predict the energy consumption of an electric bus. The objectives include the enhancement of battery management, the development of more environmentally favourable public transportation, and the enhancement of operating effectiveness. Electric vehicles are gaining popularity in smart cities due to their reduced carbon dioxide emissions and environmental benefits. Nevertheless, the frequent fluctuations in traffic, weather, passenger loading, and route factors continue to present a significant challenge in predicting the energy consumption of individual items. In order to evaluate historical operational data and propose energy consumption trends, the investigation implements sophisticated machine learning methodologies, including ANN, Random Forest, SVM, and LSTM networks. The models provided assist transportation officials in making real-time decisions and enhance prediction accuracy by recognising nonlinear interactions between the factors that influence them. Machine learning-based forecasting techniques have been shown to be highly advantageous for electric bus fleets in terms of energy conservation, optimised charging schedules, reduced operational costs, and improved fleet management consistency and consistency. The article advocates for the implementation of environmentally favourable electric mobility alternatives, as intelligent transportation technology has undergone a significant transformation.