Analysis of drivers eye of the scene". Eye tracking is a tool - - PDF document

analysis of drivers eye
SMART_READER_LITE
LIVE PREVIEW

Analysis of drivers eye of the scene". Eye tracking is a tool - - PDF document

Eye Movements and Eye Tracking 32nd ICTCT Conference in Warsaw, Poland on 24/25 October 2019 In front of a scene, a person's eyes move between points that capture their attention, " Don't wait for accidents to happen " especially


slide-1
SLIDE 1

Analysis of drivers’ eye movements ahead of pedestrian crossings: the case study of Olomouc

Warsaw, Poland

Anton Pashkevich, Matúš Šucha

32nd ICTCT Conference in Warsaw, Poland on 24/25 October 2019 "Don't wait for accidents to happen"

Eye Movements and Eye Tracking

2019-10-24 Warsaw, Poland 2

  • In front of a scene, a person's eyes move between points that capture their attention,

especially objects in movement, with which they manage to recreate a "cerebral image

  • f the scene".
  • Eye tracking is a tool to analyse the movement and behaviour of the eyes in the

presence of stimulus.

  • Eye tracking allows to know the eye behaviours:
  • It can give a possibility to analyse the fixation point of the gaze, the movement of the eyes in

relation to the head, and other types of eye movements.

  • It can give us information about human behaviour in any environment to analyse preferences of

users.

  • The standard model in eye-tracking studies is constituted by two concepts:
  • Fixation: the moment in which the eyes remain fixed on an object and it is possible to appreciate

it in detail

  • Saccades (or saccades movements): the rapid eye movements between two fixations.

Eye Tracking as a road safety research toolce

  • Eye-tracking device for road safety research
  • Well-established since 1970s.
  • Stationary and mobile devices
  • Recorded are:
  • Gaze
  • Fixation
  • Fixation duration
  • Saccade
  • The systems of capture of eye movements are becoming

very relevant today:

  • They allow us to extract objective data when a person looks

at a stimulus, in a specific situation

  • They allow us to know where the person is specifically looking

2019-10-24 Warsaw, Poland 3

Previous research work

  • Two perspectives: driver and pedestrian
  • Pedestrians:
  • Zito et al. (2015) – old/ young pedestrains on unsignalized crossing, virtual reality
  • Jiang et al. (2018) – young pedestrains on signalized crossing, outdoor
  • Kurvers (2018) – young pedestrains on unsignalized crossing, simulator
  • Biassoni et al. (2018) – children/ adults on unsignalized crossing, pictures
  • Drivers:
  • Chapman and Underwood (1998), Underwood et al. (2002), Konstantopoulos et al. (2010) –

drivers’ perception in different conditions, film video and virtual routes

  • Ābele et al. (2018) – young drivers’ perception of adult and child on unsignalized crossings,

simulator

  • Chen et al. (2019) – drivers’ recognition of pedestrian road-crossing intentions on unsignalized

crossings, video-based simulation

  • Lack of outdoor

2019-10-24 Warsaw, Poland 4

Case study plan

  • Aim:
  • General: to understand deeper drivers’ eye movements ahead of pedestrian crossings (and

to assess drivers’ behaviours towards different situations as well as relations “drivers vs.

  • ther road-users”)
  • Concrete: how do drivers scan pedestrian crossings?
  • Technical background:
  • Eye-tracking glasses: to analyze attention and focus of drivers
  • Test field:
  • Route in Olomouc (around 15 minutes driving time when taffic is free)
  • Objects - pedestrian crossings
  • Time and conditions of driving tests: October 2019, early in the morning (peak-hours)
  • Sample of drivers: 9 drivers, 22-32 ages
  • Work in progress

2019-10-24 Warsaw, Poland 5

Field test - Route

2019-10-24 Warsaw, Poland 6

  • General characteristics
  • 38 pedestrian crossings

(from drivers’ point of view):

  • 28 on mid-block section

(20 without traffic lights)

  • 6 after turning right/ left

(4 without traffci lights)

  • 4 on the roundabout
  • Pilot study:
  • C3, C4, C5
slide-2
SLIDE 2

Eye tracking glasses

  • Eye-tracking device
  • Gaze
  • Fixation
  • Fixation duration
  • Additional:
  • Distance of gazes/

fixations

  • Visibility Distance

2019-10-24 Warsaw, Poland 7

Patterns of pedestrian crossings' scanning

  • Each view was considered

separately:

  • Amount of gazes
  • Fixation and its duration
  • Distance
  • Data was „cleaned”
  • Types of patterns:
  • „Empty” pedestrian crossing
  • Pedestrian available

2019-10-24 Warsaw, Poland 8 "garbage" Gazes included After/ after (car opposite) Crossing ahead/ crossing ahead (car ahead) Before/ before (car opposite) Crossing centre/ crossing centre (car ahead)/ crossing centre (car

  • pposite)

Car ahead Crossing left/ crossing left (car

  • pposite)

Mirror wing/ mirror back Left sidewalk/ left sidewalk (car

  • pposite)

Road ahead/ road ahead (car ahead) Right sidewalk Road left Road left (car opposite) Sign right/ sign left House right/ house left 2019-10-24 Warsaw, Poland 9 2019-10-24 Warsaw, Poland 10

Patterns of pedestrian crossings' scanning – “Cleaned” Data

  • Before
  • After

2019-10-24 Warsaw, Poland 11

Patterns of pedestrian crossings' scanning – Assumptions

  • Pedestrian available:
  • Scanning from right to left
  • Way of scanning dependes on distance to pedestrian on

crossing

  • Driver fixed minimum one time on pedestrian

(experienced drivers did it less)

  • “Empty” pedestrian crossing
  • Scanning from centre of crossing
  • Scanning process is finished in the ahead part of

pedestrian crossing

  • Pedestrian crossing configuration could be important
  • Car coming from opposite lane is important
  • It „reduces” a scanning horizon
  • Amount of scanning points depends on driver’s

experience

2019-10-24 Warsaw, Poland 12

slide-3
SLIDE 3

Summary

  • We have a long way ahead of us...
  • Different scanning patterns of pedestrian crossings
  • Importance of outdoor test driving (for example, opposite cars’ incfluence)
  • Impotrance of horizontal road marking on pedestrian crossings
  • Further research by using experimental car:
  • synchronization of all devices to record
  • eye trackin glasses,
  • speed,
  • acceleration/ stop,
  • responses of ADAC,
  • Video cameras,

2019-10-24 Warsaw, Poland 13

Thank you for your attention

Anton Pashkevich Email: anton.pashkevich@upol.cz Matuš Šucha Email: matus.sucha@upol.cz

2019-10-24 Warsaw, Poland 14

32nd ICTCT Conference in Warsaw, Poland on 24/25 October 2019 "Don't wait for accidents to happen"