Driver Classification for Optimization of Energy Usage in a Vehicle

https://doi.org/10.1016/j.procs.2012.01.077Get rights and content
Under a Creative Commons license
open access

Abstract

Real time monitoring of some key dynamical parameters of a vehicle provide critical information about the driving styles and expectations of vehicle drivers. Some of these key dynamical parameters include vehicle acceleration, braking, speeding index and throttle activity index. This paper presents a simple classifier that uses the estimated values of the above parameters to classify a driver into one of three categories, aggressive, moderate and conservative. The proposed classifier is computationally more efficient compared to other conventional classifiers, such as K-nearest neighbor algorithm, and hidden Markov model. Also, it filters the reference data set in an intelligent fashion. In a dual-power vehicle, such as a hybrid electric vehicle, this kind of classifier can be used to develop an optimum shift schedule, or an optimum engine on-off strategy, and estimate the available amount of regenerative energy.

Keywords

Driver classification
K-Nearest Neighbor algorithm
powertrain signals
principal component analysis

Cited by (0)