Constructing electric city bus energy economy forecasts using CNN2D Data

Abstract

Transportation systems are becoming more and more electrified; city buses in particular have a lot of possibilities. A thorough comprehension of real-world driving information is necessary for fleet management and vehicle design. Efficient operation of alternative powertrains requires careful consideration of several technical elements. Energy demand uncertainty leads to cautious design, which means high costs and inefficiency. The intricacy and interdependence of the factors in this issue prevent both industry and academics from coming up with analytical solutions. Through streamlined processes, precise energy demand forecast allows considerable cost savings. The goal of this research is to make the energy economics for battery electric bus (BEBs) more transparent. To describe speed profiles, we propose new sets of explanatory factors that we use in our potent machine learning techniques. We create five distinct algorithms and thoroughly evaluate them in terms of accuracy of predictions, robustness, and general application. With the careful feature selection, our models performed very well, achieving a prediction precision of over 94%. Manufacturers, fleet managers, and communities have a great deal of potential to change mobility with the help of the suggested technique, opening the door for environmentally friendly public transit.