Location

Manchester Village, Vermont

Date

28-6-2017

Session

Session 5 — Poster Session B

Abstract

Two methods of time series analysis were applied to naturalistic driving data. The SAX method reduces the dimensionality of the data by discretizing and quantizing it into distinct symbols. The matrix profile method works on raw data and computes a Euclidian distance measure between subsequences of the time series. Both methods can be used to search for motifs and discords (anomalies) in the data. We discuss the applications of these methods to look for driving patterns and show an example of a left turn that was identified using both methods. After comparing the methods, the matrix profile was the preferred method.

Rights

Copyright © 2017 the author(s)

DC Citation

Proceedings of the Ninth International Driving Symposium on Human Factors in Driver Assessment, Training and Vehicle Design, June 26-29, 2017, Manchester Village, Vermont. Iowa City, IA: Public Policy Center, University of Iowa, 2017: 270-276.

Share

COinS
 
Jun 28th, 12:00 AM

Time Series Categorization of Driving Maneuvers Using Acceleration Signals

Manchester Village, Vermont

Two methods of time series analysis were applied to naturalistic driving data. The SAX method reduces the dimensionality of the data by discretizing and quantizing it into distinct symbols. The matrix profile method works on raw data and computes a Euclidian distance measure between subsequences of the time series. Both methods can be used to search for motifs and discords (anomalies) in the data. We discuss the applications of these methods to look for driving patterns and show an example of a left turn that was identified using both methods. After comparing the methods, the matrix profile was the preferred method.