Neural Adaptive Control Strategy for Hybrid Electric Vehicles with Parallel Powertrain

Many theoretical control strategies have been proposed for hybrid electric vehicles (HEVs) during the past decade. Some of these theoretical control strategies are not suitable for real-time applications mainly because of their sensitivity to vehicle parameter variations and different driving habits of the drivers. The computation times of such algorithms are also long because of their high accuracy demand. In this paper, the equivalent consumption minimization strategy (ECMS) is used and a faster solution algorithm is proposed to decrease the computation time while keeping the same accuracy. In addition, a neural adaptive network is proposed to decrease the sensitivity of the algorithm to drive cycle variations with drive cycle recognition.