Sensitivity Analysis of a Real-Time Trip Planning Assisted Energy Management System for Connected Plug-In Hybrid Electric Vehicles

Recent advances in transportation have proved that the use of upcoming trip data in energy management of plug-in hybrid electric vehicles (PHEVs) is extremely effective in promoting their fuel economy (FE). However, traffic mispredictions as well as the additional overhead of processing route data induce challenges in use of such strategy in practice. With this in mind, the present paper evaluates practical implementability of a trip planning assisted energy management system (TPAEMS). Toward the end, online integration of the TPAEMS with real-time data and also fast implementation of the trip planning algorithm is first outlined. Then, the TPAEMS under consideration is implemented for the Toyota Prius PHEV and its performance is compared to the vehicle's baseline energy management system (EMS) through model-in-the-loop simulations. In addition, hardware-in-the-loop experiments are carried out to evaluate the computational time of the TPAEMS and verify its real-time implementability. Furthermore, the sensitivity of the TPAEMS to traffic mispredictions is studied through a comprehensive sensitivity analysis. In particular, traffic prediction errors are modeled using probabilistic uncertainty modeling and also real-world data. Numerous repeated random samples of uncertainties are utilized to carry out extensive Monte Carlo simulations followed by statistical analyses of the outcomes. Our results indicate that the TPAEMS under consideration is sufficiently fast to be implemented in real time and exhibits low sensitivity to traffic mispredictions so that it maintains FE gain in real-world scenarios compared to the baseline EMS.