A Time and Energy Efficient Routing Algorithm for Electric Vehicles Based on Historical Driving Data

• Published in: IEEE transactions on intelligent vehicles 2017-12, Vol.2 (4), p.308-320
• Main Authors: Bozorgi, Amir Masoud, Farasat, Mehdi, Mahmoud, Anas
• Format: Article
• Language: English
• Subjects: Algorithm design and analysis; Classification; Classification algorithms; clustering; Clustering methods; Data mining; efficient routing; electric vehicle (EV); Electric vehicles; Energy consumption; Energy efficiency; travel time; Vehicle routing
• ISSN: 2379-8858; 2379-8904
• DOI: 10.1109/TIV.2017.2771233

A routing algorithm that leads to extended driving range and battery longevity of electric vehicles (EV) is proposed. In addition to locating the time and energy efficient routes, the proposed algorithm provides a desired speed profile to be tracked by the driver. Data mining techniques are employed for extracting the desired speed profile for the goal driver from a set of historical driving data. In order to select data with strong analogy to those of the goal driver's vehicle and driving conditions, driving and vehicle attributes are defined. The historical driving data are clustered and the class of the goal driver among clustered data is determined through classification. Eventually, the required travel time and energy consumption corresponding to historical speed profiles are evaluated and the time and/or energy efficient route along with the desired speed profile are determined. The proposed method is tested on a set of data gathered in the Warrigal project, which provides real vehicle state information. Since the consumed energy data are not available in this dataset, a detailed EV model is adopted to estimate the energy consumption. The obtained results verify the effectiveness of the proposed routing algorithm in locating the time and/or energy efficient routes.