Energy Consumption Modelling and Parameter Estimation for Autonomous Electric Shuttles

Modelling of energy consumption in electric vehicles is important for operation planning and smart charging strategies. This study proposes an energy consumption modelling and estimation strategy for autonomous electrified shuttles used for public transportation in Linköping. Operational experiences have highlighted the significant influence of weather conditions on the vehicles’ range, and the main complicating factor here is that detailed HVAC system data is unavailable. Therefore, a model is developed for estimating the energy consumption of the shuttles based on a physical formulation of vehicle dynamics and thermodynamics principles under varying weather conditions. The estimation relies on a fusion of predicted meteorological data, such as ambient temperature, and historical shuttle records, including interior temperature, vehicle speed, and discretized State of Charge (SoC). The lack of detailed HVAC operation information, e.g., cooling followed consecutively by heating within the same energy prediction profile, makes energy consumption estimation a challenging task. Even with some dynamic data available specifically for the shuttle’s short waiting time, these challenges render the HVAC system unobservable for nearly identical operations. To overcome this issue, a statistical approach combined with parameter estimation is employed to determine vehicle dynamic parameters and compute the thermal energy consumption and parametrization of the HVAC system. The data-driven model shows a satisfactory fitting level of R 2 with 0.85.