Traffic-Aware Vehicle Energy Management Strategies via Scenario-Based Optimization

This paper explores the development of traffic-aware energy management strategies by means of scenario-based optimization. This is motivated by that fact that real driving conditions are subject to uncertainty, thereby making the real-time optimization of the energy consumption of a vehicle to be a challenging problem. In order to deal with this situation, we employ the current framework of complete vehicle energy management in a receding horizon fashion, in which we consider random constraints representing realizations of exogenous signals, i.e., the uncertain driving conditions. Additionally, we study three methods for velocity prediction in energy management strategies, i.e., a method based on (average) traffic flow information, a method based on Gaussian process regression, and a method that combines both. The proposed strategy is tested with real traffic data using a case study of the power split in a series-hybrid electric vehicle. The behavior of the battery, control inputs and fuel consumption generated with the resulting strategies are compared against the optimal solution from an offline benchmark and a situation with perfect prediction of the future, For the considered case, the use of a Gaussian process regression and the traffic speed achieves near optimal fuel consumption.