Björn Peters, VTI
Active Attention
Active Attention
A driving simulator experiment
Björn Peters, VTI
Active Attention
Objective
Evaluation of an method to detect
driver fatigue impairment based on
analysis of the driver’s micro control
Björn Peters, VTI
Active Attention
Method
10 shift workers ( 5 male, 5 female )
• 2 hours drive, two occasions (morning)
– after a night’s work – after a normal sleep
• Rural road, oncoming cars
• Double blind test
– Data in 5 minute blocks, random order, missing
data replaced – equal number of sections for all
subjects
Björn Peters, VTI
Active Attention
Subjects
• Mean age 37 years
• Mean license age 16.4 years
• Mean annual driving distance 9500 km
• 4 had near accident experiences
• Most shift workers for 5 - 9 years
Björn Peters, VTI
Active Attention
Measures
Driving behaviour Physiology
– speed - EOG ( duration, freq .) – lateral position
– TLC KSS ratings
– steering wheel angle - driver (5 min)
– lateral acceleration - test leader (5 min) – steering wheel torque
Questionnaires
- pre and post
Björn Peters, VTI
Active Attention
Road – vehicle geometry
Björn Peters, VTI
Active Attention
Analysis approach
• Comparing AA results to KSS driver ratings
– correlation
– multiple regression
• Reference
– Driver’s lateral and longitudinal control
– Blink duration (< 0.5 s) and frequency
– Test leader KSS ratings
Björn Peters, VTI
Active Attention
KSS Driver Ratings
Average KSS (1 - 9)
1 2 3 4 5 6 7 8 9 10
300
900
1500
2100
2700
3300 3900
4500
5100
5700
6300
6900 Driving time (s)
KSS Average KSS Alert
Average KSS Fatigued
Björn Peters, VTI
Active Attention
KSS Driver and Test Leader
Average KSS (1 - 9)
1 2 3 4 5 6 7 8 9 10
Driving time (s)
KSS Average Driver KSS Alert
Average Driver KSS Fatigued Average TL KSS Alert
Average TL KSS Fatigued
Björn Peters, VTI
Active Attention
Results - driving behaviour
• 4 drivers were not able to complete fatigued driving condition
• 11 “accidents”
– 4 wheels off the road – 10 “fatigued”, 1 “alert”
– 10 at KSS level 9, 1 at KSS level 8
•
Björn Peters, VTI
Active Attention
Correlations – all drivers
KSS – Driving behaviour
Björn Peters, VTI
Active Attention
Correlations per driver
KSS – Driving behaviour
Björn Peters, VTI
Active Attention
Regression Analysis Driver 1
Björn Peters, VTI
Active Attention
Correlations – all drivers
KSS to AA0 -5
Björn Peters, VTI
Active Attention
Correlations per driver
KSS to AA0 -5
Björn Peters, VTI
Active Attention
Driver 1
Björn Peters, VTI
Active Attention
Driver 3
Stopped
driving
Björn Peters, VTI
Active Attention
Driver 8
Björn Peters, VTI
Active Attention
Driver 1 block 3
Björn Peters, VTI
Active Attention
Driver 1 – block 36
Björn Peters, VTI
Active Attention
Steering angle Driver 1
Björn Peters, VTI
Active Attention
Road departures
4 wheels off the road
Björn Peters, VTI
Active Attention
Subject 3 – block 34
Björn Peters, VTI
