Driving pattern prediction for an energy management system of hybrid electric vehicles in a specific driving course

Energy management systems that consider a whole driving pattern have been proposed to improve the fuel consumption and the emissions of hybrid electric vehicles. In this research, we propose a new energy management system that includes a driving pattern prediction system. The proposed method pay attention to a commuting to work because a commuting occupies the large portion of mileage and are repeated day-to-day. Therefore the database of driving patterns can be simply constructed by the proposed clustering method, and the future driving pattern can be predicted from the database. In this paper, it is shown that the actual driving pattern is changed by external factors, and we propose an improve method of driving pattern prediction. The experimental result shows the availability of the proposed system.