Data mining approach for range prediction of electric vehicle

• Autor(es): Ferreira, João C. ; Monteiro, Vítor ; Afonso, João L.
• Data: 2012
• Identificador Persistente: http://hdl.handle.net/1822/19439
• Origem: RepositóriUM - Universidade do Minho
• Assunto(s): Range prediction; Data mining; Driver profile; Range anxiety problem; Electric vehicle
• Tipo de Documento: Documento de conferência
• Idioma: Inglês

Descrição

Our work proposal is based on the past driving data that are stored in a driver profile, and using real time information about the Electric Vehicle parameters (e.g. speed and energy stored in the batteries), combined with external parameters (e.g. condi-tions of roads, traffic, and weather), determine the range autonomy accurately, taking into account the historical driver behavior. The driver profile is based on the stored data, which acts as training set for a Data Mining approach, in order to estimate the Electric Vehicle range. The Data Mining approach uses a regression model aiming to find the better range autonomy, which is used to represent the current Electric Vehicle range autonomy on a map.