The Detection of Driver Cognitive Distraction Using Data Mining Methods

1
The Detection of Driver Cognitive Distraction
Using Data Mining Methods

• Presenter Yulan Liang
• Department of Mechanical and Industrial
  Engineering
• The University of Iowa

2
Driver distraction

• Driver distraction and inattention has become a
  leading cause of motor-vehicle crashes
• Nearly 80 of crashes and 65 of near-crashes
  (the 100-car study)
• Increasing use of In-Vehicle Information Systems
  (IVISs), such as, navigation systems, MP3
  players, and internet services.
• Driver distraction represent a big challenge for
  developing IVISs
• Benefits of the IVIS functions
• Safety
• One solution driver distraction mitigation
  systems

People use In-Vehicle Information Systems (IVISs)
during driving
3
Driver distraction mitigation systems

• Distraction detection is a crucial function
• Cognitive distraction
• Visual/manual distraction
• Simultaneous(dual) distraction

• Indicators of distraction
• Detection techniques

An overview of driver distraction mitigation
systems
4
Indicators of driver distraction

• Cognitive distraction (subtle, no direct measures
  of mind off road)
• Concentrate gaze distribution
• Impair information consolidation
• Degrade driving performance (less serious and
  consistent)
• Impair driver adaptation in tactical driving

Suitable for real-time detection
Performance indicators
--Eye gaze Duration and location of
fixations Distance of saccades Duration,
location, distance, and speed of smooth pursuits

• --Driving performance (less serious and
  consistent)
• Abrupt steering control
• Large lane-position variability
• Miss safety-critical events

Not suitable for real-time detection
5
Detection algorithm for driver distraction

• Driving is complex and continuous human behavior
• Data mining approaches are suitable to detect
  driver distraction
• Insufficient knowledge impedes using theories to
  detect distraction precisely
• Data mining techniques can detect non-linear and
  time-dependent relationships
• 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)

6
Bayesian Networks (BNs)
Cognitive distraction

• To model probabilistic relationship among
  variables
• wide applications, especially modeling human
  behavior
• Three kinds of variables
• Hypothesis, evidence, hidden
• Conditional dependency

Eye movement pattern

Bayesian Networks (BNs)
Eye movementsDriving performance
7
Static and Dynamic BNs

• Static BNs (SBNs)
• in single time point
• Dynamic BNs (DBNs)
• across time (Markov process)
• Comparison btw SVM and BNs
• Both can model complex relationships
• Results of BNs can quantify relationships using
  information theory measures (such as mutual
  information)
• DBNs can model time-dependent relationship
• SVMs are more computational efficient than BNs.

A dynamic BN
8
Methods

• Data source
• two cognitive conditions
• auditory stock ticker tracking the change and
  overall trends of two stock prices
• without visual distractors
• 4 IVIS drives and 2 baseline drives (15 minutes
  each)
• to define distraction for models
• data collection (60Hz)
• eye movements
• gaze screen intersection coordinates
• Driving performance
• lane and steering position

Driving scenario
9
Data reduction
Plot of eye data

• Eye movements
• eye data ?eye movements
• 7 eye movement measures
• 3 driving performance measures
• lane position
• steer wheel position
• steering error

fixation
-duration-position
smooth pursuit
-duration-distance -speed -direction
blink frequency
10
Training Data

• Summarization
• window size
• (5, 10, 15, or 30 s)
• Training data
• SBNs SVMs
• DBNs
• 2/3 of total data

(19 measures)
11
SVM and BN training parameters

• SVMs
• Radial Basis Function (RBF)
• 10-fold-cross-validation to obtain C and ? in the
  range of 2-5 to 25
• Continuous predictors (performance measures)
• LIBSVM Matlab toolbox
• BNs
• No hidden node and constrained network structure
• Training sequences for DBN 120 seconds long
• Discrete predictors
• a Matlab toolbox (Murphy) and an accompanying
  structural learning package (LeRay)

12
Using SVMs and DBNs to detect cognitive
distraction
SVM prediction for a participant
d'
Comparison between BNs and SVMs
13

• Changes in drivers eye movements and driving
  performance over time are important predictors of
  cognitive distraction.
• SVMs have some advantages over SBNs
• Parameter selection 10-fold across-validation
• Computational ease training time
• Improving algorithm
• Consider time-dependent relationship in behavior
• Reduce computational load

14
A layered algorithm to detect cognitive
distraction

• Off-line supervised clustering identifies
  multiple feature behavior based on subset of
  behavioral measures based on the training data
• Temporal eye movement measures
• Spatial eye movement measures
• Driving performance measures
• The higher layer DBNs identify cognitive state
  from the feature behavior(cluster labels) with
  consideration of time dependency

Different from clustering, supervised clustering
more likely produce meaningful clusters in terms
of driver cognitive state.
15
Supervised clustering

• categorize classified data

The fitness function of supervised clustering
(Zeidat et al., 2006) X is a clustering
solution, ß is the parameter to balance the ratio
of impurity and penalty in the fitness function,
k is the number of clusters in X, n is the total
number of data, and c is the number of classes in
the data.
16
Supervised clustering algorithm

• Single Representative Insertion/Deletion Steepest
  Decent Hill Climbing with Randomized Restart
  repeat something similar to SPAM r times and
  chose the best
• REPEAT r TIMES
• curr a randomly created set of representatives
  (with size between c1 and c)
• WHILE not done DO
• Create new solution S by adding a
  non-representative or removing a representative
  in curr (if size(curr) k, new possible
  solutions are in size of k1 and k-1 )
• Determine the element s and S for which the
  objective function in SPAM q(s) is minimal (if
  there is more than one minimal element, randomly
  pick one)
• IF q(s)ltq(curr) THEN currsELSE IF q(s)q(curr)
  AND sgtcurr THEN currsELSE terminate and
  return curr as the solution for this run
• Report the best out of the r solutions found

17

• Thank you !!
• Questions ??