Personalized Prediction of Driving Energy Consumption based on Participatory Sensing

The advent of abundant on-board sensors and electronic devices in vehicles populates the paradigm of participatory sensing for intelligent transportation systems, by harnessing crowd-sourced data gathering to enable knowledge discovery in various applications, such as distance-to-empty prediction and eco-routing. While participatory sensing can provide various driving and vehicle data, there lacks a systematic study on effectively utilizing the data for personalized prediction applications. There are considerable challenges on how to interpolate the missing data from a sparse dataset, which often arises from participatory sensing. This paper presents and compares various personalized prediction approaches for driving energy consumption, including a blackbox approach that identifies driver/vehicle/environment-dependent factors and a collaborative filtering approach that uses matrix factorization. To evaluate the effectiveness of our approaches, a case study of distance-to-empty prediction for electric vehicles is conducted based on the participatory sensing data. Our approaches are shown to effectively improve the prediction accuracy.