Automobile Injury Prevention with Eye Tracking
trial one
Gaze ratio is performing accurately when detecting change in gaze direction
trial two
Eye width to height ratio is performing accurately when detecting blinking
background
Most driver alert programs on the market are based off of steering wheel trackers
These trackers focus on the steering angles and steering movements
They alert the driver when the steering wheel movements become out of the normal
The programs are designed to prevent drowsy driving by using the steering wheel to track that
about eyesight
The average person blinks 10 times per minute.
The eye blinks less when focused
Focusing for a long time can lead to dry and tired eyes
Blinking increases when fatigued
steering wheel vs. eye tracking
Using the steering wheel to track drowsy and distracted driving makes it harder to see early signs of drowsiness and distraction and alert the driver to take a break.
Using eye tracking makes it easier to see early signs of fatigue and alert the driver before the steering is affected.
Goals
To help prevent drowsy driving as well as distracted driving accidents.
Distracted driving accounts for 8.5% of all car crashes in 2019. Distracted driving accidents rises each year.
If the vision of the driver is not on the road for the specified time (seconds), the sensor will send an audible message to the driver to bring their attention back.
The time is not specified because it depends on the driving speed.
These alerts can prevent car accidents from fatigue and make it much safer for everyone on the road.
code
We have designed a sophisticated code using Python to simulate a realistic driving environment to present the benefits of eye tracking in automobile safety. Our coding adds distractions to the subject operating the driving simulator to test whether or not eye tracking software will overcome such distractions.
Note: the gaze ratio varies depending on the video/subject, as well as the lighting of the scene
The code for eye tracking is below:
https://drive.google.com/file/d/1o8-zOcV_Fj5eAg2lZ9pTtS1dHXUsfajR/view?usp=sharing
https://drive.google.com/open?id=1qca4XY8mwkGFiz-ZNrqd7kRSlLuELRCZ
Survey
Prior to any study, subjects are asked a series of questions to ensure that there is no bias in this study.
Age: younger drivers are more likely than older drivers to be involved in automobile collisions
Sex: male drivers are more likely than female drivers to be involved in automobile collisions
Race: according to 2006 traffic data, 58% of traffic fatalities were White, while 13% were Hispanic, and 11% were Black
Traffic citations: a driving record with many traffic citations indicates unsafe driving
Automobile collision: the amount of automobile collisions a driver is involved in may indicate their behavior in driving
Red light tickets: a greater amount of red light tickets may indicate careless or dangerous driving
Written/verbal warnings: warnings given may indicate a driver was driving recklessly, albeit possible unique situations
Miles driven per day: a frequent driver may be at greater risk, which is related to the amount of collisions involved in
Hours driven per day: hours driven relates to miles driven, as it indicates traffic levels
Time of day when driving: visibility is decreased at night, traffic is reduced in the middle of the day
Ride hailing: drivers that Uber or Lyft are generally safer drivers
Limitations
No access to the lab to conduct the experiment
Number of participants
Variability in age and gender of the participants
Area of subjects
Demographics
No calibration of the eye movement video
Post Study
Subjects are asked to rate the following statements with more/less/same:
Difficulty in focusing on the road in simulator compared to real driving?
Difficulty in seeing the road from the simulator?
Attentiveness while driving with simulator in comparison?
Noise level while driving with simulator compared to real driving?
Subjects are also asked the following questions:
Did eye tracking help you maintain focus?
Do you believe that eye tracking makes driving safer?
Covid19's Impact
Made it impossible to run trials involving human participants
No access to eye tracker or similar devices
The proposed project was needed access to technology that is not available at home
Project Team
Rosalie Beirne
Undergraduate IE Major
Eye tracking interested me because I like the possibility for that to be used in disability devices and to prevent injuries
Anas Sayuory
Graduate IE student
Yilong Wang
Sandra Ishwait