DRIVING PATTEN PREDICTION BASED ON DATA DRIVEN METHOD - A SAMPLE IN CITY BUSES

摘要

Driving pattern is one vital impact factor to the vehicle energy consumption. In this paper, based on one-month operation data of a city bus, the velocity profiles are normalized to minimum 10 and maximum 15 driving patterns in a reasonable operation period. Markov-chain model is suggested to generate a short-term predicted driving mode, which includes several normalized driving-patterns. This accurate prediction can be used to estimate the remaining mileage of battery electric vehicles (BEVs) and reduce the fuel consumption of hybrid electrical vehicles (HEVs).