An Advanced Micro Energy Consumption Model for Automated Vehicle by Empirical Data

Abstract: While the Adaptive Cruise Control (ACC) system in automated vehicles (AVs) is expected to impact transportation energy significantly, existing ACC vehicle energy consumption models only directly adopt those developed with Human-driven Vehicle (HV) data without even slight adaptation or even calibration to accommodate unique ACC vehicle energy consumption features. This study will investigate how accurately HV data-based models can predict the energy consumption of ACC vehicles. Empirical trajectory data and corresponding instantaneous energy consumption rates from both ACC vehicles and HVs were collected. We adopted two classical HV data-based models to fit these data. The calibration results indicated that these models yield only around 20\~30\% prediction errors for ACC vehicles. To further improve the prediction accuracy, this study designed an AA-Micro model by incorporating components of multiple classic energy consumption models that better capture ACC energy consumption features (e.g., heterogeneous piece-wise driving behavior). With this, the AA-Micro model achieves lower than 10\% prediction errors. The AA-Micro model’s high consistency across different test runs was verified with statistical significance tests, demonstrating its adaptability in different driving profiles. To confirm the discrepancies between the energy consumption features of ACC vehicle HVs, more statistical significance tests were conducted to show that the AA-Micro model cannot be directly applied to HV data. The findings by calibrated AA-Micro models revealed that ACC vehicles consume approximately 70-120 J more energy than HVs for each meter traveled, indicating that there is still some room for AVs to learn from human drivers. The development and validation of the AA-Micro model offer a more accurate tool for predicting AV energy consumption, addressing the limitations of existing HV-based models and highlighting the distinct energy consumption patterns of ACC systems.