Home // ICAS 2017, The Thirteenth International Conference on Autonomic and Autonomous Systems // View article
Authors:
Igor Vasconcelos
Rafael Vasconcelos
Bruno Olivieri
Markus Endler
Methanias Júnior
Keywords: Online Outlier Detection; Complex Event Processing; In-Vehicle Sensing; Online Driving Behavior Detection; Smartphone
Abstract:
This paper presents an approach for online outlier detection over multiple data streams based on Complex Event Processing (CEP) to enable driving behavior classification. Driving is a daily task that allows people to move around faster and more comfortably. However, more than half of fatal crashes are related to recklessness behaviors. Reckless maneuvers can be detected with good accuracy by analyzing data relating to the driver-vehicle interaction, abrupt turnings, acceleration and deceleration, for instance. In this paper, we investigate if off-the-shelf smartphones can be used to an online detection of driving behavior. To do so, we have adapted the Z-Score algorithm, a classical outlier detection algorithm, to perform online outlier detection as a data stream processing model, which receives the smartphone and in-vehicle sensors data as input. The evaluation of the approach was carried out in a case study to assess the algorithm. Our results indicate that the algorithm’s performance is fairly good in a real world case study since the algorithm’s accuracy was 84% and the average processing time was 100 milliseconds.
Pages: 73 to 78
Copyright: Copyright (c) IARIA, 2017
Publication date: May 21, 2017
Published in: conference
ISSN: 2308-3913
ISBN: 978-1-61208-555-5
Location: Barcelona, Spain
Dates: from May 21, 2017 to May 25, 2017