There is accumulating evidence that driver distrac- tion is a leading cause of vehicle crashes and incidents. In par- ticular, increased use of so-called in-vehicle information systems (IVIS) have raised important and growing safety concerns. Thus, detecting the driver’s state is of paramount importance, to adapt IVIS, therefore avoiding or mitigating their possible negative effects. The purpose of this presentation is to show a method for the nonintrusive and real- time detection of visual distraction, using vehicle dynamics data and without using the eye-tracker data as inputs to classifiers. Specifically, we present and compare different models that are based on well-known machine learning (ML) methods. Data for training the models were collected using a static driving simulator, with real human subjects performing a specific secondary task [i.e., a surrogate visual research task (SURT)] while driving. Different training methods, model characteristics, and feature selection criteria have been compared. Based on our results, using a BSN has outperformed all the other ML methods, providing the highest classification rate for most of the subjects. Index Terms—Accident prevention, artificial intelligence and machine learning (ML), driver distraction and inattention, intel- ligent supporting systems.

1. Real-Time Detection System of
Driver Distraction
Using Machine Learning

2. Contents…
 Machine Learning
 Introduction
 Driver Distraction
 Driver Distraction Mitigation
 Detection Algorithm
 Methods
 Advantages / Disadvantages
 Applications

3. MACHINE LEARNING
 Machine learning is a scientific
discipline concerned with the design
and development of algorithms that
allow machines to mimic human
intelligence.

4. Introduction
 80% of crashes and 65% of near crashes
involved some sort of driver distraction.
 Teens are 4x more likely to be in a wreck than
drivers over age 30.
 Motor vehicle crashes are the leading cause of
death for 16-20 year olds.

5. Driver Distraction
• Driver distraction and
inattention has become a leading
cause of motor-vehicle crashes
 IVIS
• Driver distraction represent a big
challenge for developing IVISs

6. Types of Distractions:
There are 3 types of distractions:
Visual Distractions: Anything that
takes your eyes off the road.
Manual Distractions: Anything that takes
your hands off the steering wheel.
Cognitive Distractions: Anything that takes
your mind off driving.

7. Distractions:
 All distractions can be dangerous and life threatening
but texting is one of the most dangerous distractions because
it involves all three types of distractions.
 Other distractive activities include:
» Using a cell phone
» Eating and drinking
» Talking to passengers
» Grooming
» Reading, including map
» Using PDA or navigation system
» Watching a video
» Changing the radio station, CD,
 Mp3 player or other device

8. How Cell Phones Distract
 Visual – Eyes off road
 Mechanical – Hands off wheel
 Cognitive – Mind off driving
 CHALLENGE: Drivers don’t
understand or realize that talking
on a cell phone distracts the brain
and takes focus away from the
primary task of driving.

9. Sleepy Driving

10. Sleepy Driving…
 100,000 reported crashes per year are as a result
of drowsiness. 1,500 of them result in deaths.
 55% of those crashes were caused by drivers
under the age of 25.

11. Driver distraction mitigation
systems
 Distraction detection is a crucial
function
oCognitive distraction
oVisual/manual distraction
oSimultaneous(dual) distraction

12. Driver state-----------------
· Physiological responses
· eye glances
· fixations, saccades, and
smooth pursuits
...
Driver input-----------------
· Steer
· Throttle
· Brake
...
Vehicle state---------------
· Lane position
· Acceleration
· Speed
...
Visual/Manual
distraction
Cognitive
distraction
Model-based Driver
Distraction Detection
Mitigation
strategy
Focus of dissertation
Sensor
Techology
Mitigation
System
Strategy n
Strategy 2
Strategy 1
.
.
.
An overview of driver distraction mitigation
systems

13. Detection algorithm for driver
distraction
• Driving is complex and continuous human behavior
• Machine learning approaches are suitable to detect
driver distraction
o Linear regression, decision tree, Support Vector Machines
(SVMs), and Bayesian Networks (BNs) have been used to
identify various distractions
Support Vector Machines (SVMs)
Bayesian Networks (BNs)

14. Bayesian Networks (BNs)
 To model probabilistic
relationship among
variables
◦ wide applications, especially
modeling human behavior
 Three kinds of variables
◦ Hypothesis, evidence, hidden
H
E3E2E1
S
Bayesian Networks (BNs)
Cognitive
distraction
Eye movements
Driving performance
Eye
movement
pattern

15. Methods
SURT(surrogate visual research
task) display on the right part of
the driving simulator cockpit

16. Advantages…
 Intelligent Decisions
 Self modifying
 Multiple iterations
 Method implementation
 Use of different methods
 Test in diverse conditions.
Disdvantages…

17. Applications
 Computer vision : design and
implementation of algorithms that can
automatically process visual data.
 Information retrieval : Technologies in
order to help solve complex and
challenging business problems.
 Robot locomotion : Capabilities for robots
to decide how, when, and where to move

18. REMEMBER…
The life you save could be your own!

19. Conclusion..
 The most ambitious goals of automatic learning
systems is to mimic the learning capability of
humans, and the capability of humans to drive is
widely based on experience, particularly on the
possibility of learning from experience. ML
approaches can outperform the traditional
analytical methods. Moreover, a human’s mental
and physical driving behavior is non
deterministic.

20. QUESTIONS...

21. THANKYOU