[4I2-GS-2] Machine learning: Living with AI

[4I2-GS-2-04] Prediction of Departure and Travel Time of Individual Vehicle Based on Past Usage Patterns by Using Support Vector Machine

〇Takuya Shimada¹, Takahiro Nishigaki¹, Takashi Onoda¹ (1. Aoyama Gakuin University)

Keywords:Machine Learning, Support Vector Machine, Energy Management System, Electric Vehicle

In recent years, with concerns about environmental problems, the introduction of an Energy Management System (EMS) is progressing worldwide. EMS requires a storage battery to compensate for charging and discharging immediately. Therefore, the introduction of storage batteries mounted in vehicles into EMS is progressing. By predicting future behavior patterns of vehicles, the batteries in the vehicles can be used as city energy. This study used Support Vector Machine to predict vehicles behavior patterns one day ahead based on past vehicles behavior patterns. The conventional method to be compared is a prediction method using a Markov model. It was confirmed that TPR and ACC exceeded that of the conventional method.