**6. Conclusions**

This paper proposes a multiobjective off-line optimization-driven ACC approach for car-following automated driving scenarios that can flexibly adapt to different powertrain categories. A CV powertrain, a BEV powertrain, a parallel P2 HEV powertrain and a PS HEV powertrain are considered as test cases, and their numerical model was presented along with the related onboard control strategies. The optimal problem for car-following driving has then been outlined. Propelling energy-saving and passenger comfort improvement were selected as the two optimization targets when controlling the following vehicle's longitudinal acceleration throughout a given drive cycle. Dedicated constraints were integrated for the maximum and minimum achievable values of IVD, along with specific powertrain-related constraints. An optimization-driven control solution for the presented car-following driving problem can be obtained by implementing a DP technique. Simulation results obtained in different driving conditions highlight the potential of the proposed ACC approach in identifying improved control solutions for the following vehicle in terms of energy-saving and passenger comfort considering a wide range of powertrain categories. Up to 22.1% energy-saving and up to 48.2% reduction in the RMS of the vehicle acceleration were demonstrated by the following-vehicle led using the proposed approach compared with the preceding vehicle, depending on the tuning performed for the two optimization targets.

In general, the illustrated approach preserves the engineer's freedom to select the weights for energy-saving and passenger comfort improvement for the following vehicle's operation. The obtained optimization-driven results might be used to benchmark different ACC approaches in this way. Moreover, the proposed approach could pave the way for developing real-time-capable control algorithms for the following vehicle in car-following scenarios that mimic optimal control actions forecasted by the introduced off-line optimization-driven approach. Furthermore, improving the fidelity level for the modeling approach might be achieved in terms of powertrain, vehicle dynamics, and ACC sensing using radar, LIDAR or cameras. For example, adaptations in the onboard control logic for gear-shifting, ICE activation and power split could be examined to further

enhance the powertrain efficiency when traveling as a following vehicle in car-following driving. Finally, the optimization-drive approach could be extended considering multiple ACC-enabled vehicles traveling behind the preceding vehicle.


**Table 3.** Results for the preceding and following vehicles.

**Funding:** This research was funded by the Doctoral School of Politecnico di Torino. **Conflicts of Interest:** The author declares no conflict of interest.

#### **Abbreviations**


#### **Appendix A. Optimal V2V Driving Pareto Fronts**

**Figure A1.** Pareto fronts for fuel consumption and RMS of the vehicle acceleration in WLTP (**a**), UDDS (**b**), HWFET (**c**) and US06 (**d**)—CV powertrain.

**Figure A2.** Pareto fronts for fuel consumption and RMS of the vehicle acceleration in WLTP (**a**), UDDS (**b**), HWFET (**c**) and US06 (**d**)—BEV powertrain.

**Figure A3.** Pareto fronts for fuel consumption and RMS of the vehicle acceleration in WLTP (**a**), UDDS (**b**), HWFET (**c**) and US06 (**d**)—P2 HEV powertrain.

**Figure A4.** Pareto fronts for fuel consumption and RMS of the vehicle acceleration in WLTP (**a**), UDDS (**b**), HWFET (**c**) and US06 (**d**)—PS HEV powertrain.

**Appendix B. Time-Series of Suboptimal Control Solutions in WLTP**

**Figure A5.** Time-series for the simulation results of the CV powertrain in WLTP both as the preceding vehicle and the following vehicle for the "Opt\_Energy" suboptimal control solution.

**Figure A6.** Time-series for the simulation results of the CV powertrain in WLTP both as the preceding vehicle and the following vehicle for the "Opt\_Comfort" suboptimal control solution.

**Figure A7.** Time-series for the simulation results of the BEV powertrain in WLTP both as the preceding vehicle and the following vehicle for the "Opt\_Energy" suboptimal control solution.

**Figure A8.** Time-series for the simulation results of the BEV powertrain in WLTP both as the preceding vehicle and the following vehicle for the "Opt\_Comfort" suboptimal control solution.

**Figure A9.** Time-series for the simulation results of the P2 HEV powertrain in WLTP both as the preceding vehicle and the following vehicle for the "Opt\_Energy" suboptimal control solution.

**Figure A10.** Time-series for the simulation results of the P2 HEV powertrain in WLTP both as the preceding vehicle and the following vehicle for the "Opt\_Comfort" suboptimal control solution.

**Figure A11.** Time-series for the simulation results of the PS HEV powertrain in WLTP both as the preceding vehicle and the following vehicle for the "Opt\_Energy" suboptimal control solution.

**Figure A12.** Time-series for the simulation results of the PS HEV powertrain in WLTP both as the preceding vehicle and the following vehicle for the "Opt\_Comfort" suboptimal control solution.

#### **References**

