Location

Big Sky, Montana

Date

23-6-2009

Session

Session 3 – Poster Session A

Abstract

This paper examined a steering behavior based fatigue monitoring system. The advantages of using steering behavior for detecting fatigue are that these systems measure continuously, cheaply, non-intrusively, and robustly even under extremely demanding environmental conditions. The expected fatigue induced changes in steering behavior are a pattern of slow drifting and fast corrective counter steering. Using advanced signal processing procedures for feature extraction, we computed 3 feature set in the time, frequency and state space domain (a total number of 1251 features) to capture fatigue impaired steering patterns. Each feature set was separately fed into 5 machine learning methods (e.g. Support Vector Machine, K-Nearest Neighbor). The outputs of each single classifier were combined to an ensemble classification value. Finally we combined the ensemble values of 3 feature subsets to a of meta-ensemble classification value. To validate the steering behavior analysis, driving samples are taken from a driving simulator during a sleep deprivation study (N=12). We yielded a recognition rate of 86.1% in classifying slight from strong fatigue.

Rights

Copyright © 2009 the author(s)

DC Citation

Proceedings of the Fifth International Driving Symposium on Human Factors in Driver Assessment, Training and Vehicle Design, June 22-25, 2009, Big Sky, Montana. Iowa City, IA: Public Policy Center, University of Iowa, 2009: 118-124.

Share

COinS
 
Jun 23rd, 12:00 AM

Steering Wheel Behavior Based Estimation of Fatigue

Big Sky, Montana

This paper examined a steering behavior based fatigue monitoring system. The advantages of using steering behavior for detecting fatigue are that these systems measure continuously, cheaply, non-intrusively, and robustly even under extremely demanding environmental conditions. The expected fatigue induced changes in steering behavior are a pattern of slow drifting and fast corrective counter steering. Using advanced signal processing procedures for feature extraction, we computed 3 feature set in the time, frequency and state space domain (a total number of 1251 features) to capture fatigue impaired steering patterns. Each feature set was separately fed into 5 machine learning methods (e.g. Support Vector Machine, K-Nearest Neighbor). The outputs of each single classifier were combined to an ensemble classification value. Finally we combined the ensemble values of 3 feature subsets to a of meta-ensemble classification value. To validate the steering behavior analysis, driving samples are taken from a driving simulator during a sleep deprivation study (N=12). We yielded a recognition rate of 86.1% in classifying slight from strong fatigue.