Energy-saving control of intelligent connected plug-in hybrid electric vehicle via fusing driving intention of front vehicle

Abstract

Velocity variation of front vehicle substantially influences the driving performance and energy consumption of the following vehicle, particularly for hybrid electric vehicles (HEVs). Fusing the intention identification of front vehicle into the control of following HEV can facilitate speed optimization, energy saving and operation efficiency promotion. Inspired by this, an effective speed optimization and energy management approach is designed for the following plug-in HEV (PHEV) with the support of front vehicle’s driving style identification. To this end, an improved K-means clustering approach and the support vector machine algorithm are respectively employed to cluster and distinguish the driving intention. Next, the next step velocity of front vehicle is predicted by Elman neural network, based on the current velocity and the identified driving intention. Subsequently, a multi-objective speed optimization problem is formulated with the consideration of powertrain efficiency, vehicle comfort, tracking capability and vehicle safety. Then, a power distribution strategy is designed based on model predictive control with a clipped double Q-learning algorithm to allocate the energy flow in different energy sources. The simulation findings demonstrate that the proposed strategy not only achieves preferable following effect and comfort, but also leads to 97.01% energy economy optimality supplied by dynamic programming. • Driving intention classification according to the improved K-means method • Front vehicle’s velocity prediction by fusing driving intention identification • Velocity planning considering front vehicle’s speed and vehicle’s dynamics • Preferable energy economy of the vehicle resulted by the clipped double Q-learning

Topics & Concepts