Predictive Co-Optimization of Speed Planning and Powertrain Energy Management for Electric Vehicles Driving in Traffic Scenarios: Combining Strengths of Simultaneous and Hierarchical Methods

Adapting to the instantaneous disturbance in the driving environment and balancing the optimality and computational efficiency of control algorithms are two major challenges for the integrated optimization of active speed planning and powertrain energy management strategy. In this study, utilizing the framework of model predictive control, a simultaneous method (SM) and a hierarchical method (HM) are developed to serve as benchmarks for control optimality and computational efficiency, respectively. Then, by modifying the decoupling strategy of the HM, this study ultimately proposes a modified HM which achieves similar control effectiveness in energy saving as that of the SM and preserves high computational efficiency. The comparative validation demonstrates that, due to fierce acceleration/deceleration operation caused by the heuristic decoupling strategy adopted in the HM, the energy consumption provided by the HM is 221.5% in traffic flow scenarios (and 633.5% in the manually designed scenario) of that generated by the SM. However, by adopting the soft constraint on acceleration magnitudes, the modified HM narrows the sub-optimality in energy consumption to 3.95% compared with the SM, and it also realizes a 55.81% improvement in computation efficiency compared with the original HM.