A robust range prediction system is proposed that estimates the range of an electric vehicle (EV) during operation, using a range prediction model trained on data relevant to the energy usage of the vehicle. The range prediction model may take a greater number of inputs than may be considered by an OEM range estimator provided with the EV. The predicted range generated by the range prediction system may then be compared with the estimated range of the OEM range estimator, and/or combined into an integrated display to provide the driver with a confirmation of the estimated range, which may decrease a range anxiety of the driver. Additionally, the proposed range prediction system may generate recommendations for how to maximize (e.g., maintain or increase) the predicted range by adjusting the operation of the vehicle or coordinating operation of the vehicle with other vehicles, traffic signals, etc.
