Using Machine Learning to Simulate Electric Vehicle Efficiency in Virtual Assessments

Abstract

As the global transport landscape undergoes significant transformation, the sociological and economic implications continue to play a pivotal role. While advanced regions of the world transition towards electric mobility, introducing cleaner technologies and sustainable practices, less developed countries are confronted with multifaceted challenges. These include not only shifts in mobility options but also profound impacts on basic infrastructure, environmental landscapes, and cultural behaviours. Such complexities contribute to a slower adoption rate of electric mobility, particularly in Sub-Saharan Africa, where systemic barriers and underdevelopment further complicate the transition. This disparity underscores the pressing need for comprehensive strategies sensitive to the distinct socio-economic contexts of underdeveloped regions. It also highlights the importance of international cooperation and local policy adaptations that could facilitate a smoother transition to electric mobility, thereby enhancing the adoption rate and ensuring that the benefits of electric mobility—such as reduced emissions, improved public health, and energy independence—are realised across diverse global contexts. This article introduces a novel approach utilising a Feed Forward Neural Network machine learning model to digitally pilot electric vehicles on specific routes, without requiring physical deployment. This machine learning solution provides a costeffective method for assessing the efficiency of electric vehicles over various routes, facilitating data-driven decisions to overcome adoption barriers. The digital pilot model offers a strategic tool that allows for the simulation of electric vehicle performance under diverse operational conditions, thereby enabling policymakers and fleet operators to evaluate potential investments and infrastructure needs with greater accuracy and less financial risk. This contribution paves the way for accelerated and informed adoption of electric mobility in regions that face the greatest challenges.

Keywords: Electric vehicle, Paratransit, Vehicle charging, Vehicle-time, Minibus taxi, Machine learning, Digital pilot, Retrofit