L5- Gaze & Driving

Question 1

Outline the components of Salvucci & Gray’s (2004) 2 level control strategy for driving

Answer 1

a). Keep the appropriate distance from the road edge - use a ‘near’ point/region (closed loop)
b). Match the curvature of the road - use a ‘far’ point/region (open loop)

Question 2

Outline Land & Horwood’s (1995) study

Answer 2

-Manipulated near, middle, and far road regions
-Only far road info - matched curvature but not lane position
-Only middle - Reasonably good for both
-Only near - maintain lane position but didnt match curvature

Question 3

Evaluate Land & Horwood (1995)

Answer 3

-No retinal flow present - may exaggerate effect
-Low update rate (7Hz)

Question 4

Outline Cloete & Wallis (2011)

Answer 4

-Refresh rate (7.2Hz vs 72Hz) affected performance
-Some evidence for 2-point model but not compelling
-Likely that we use near road info to maintain lane position.

Question 5

Outline the Tangent Point Model (Land & Lee, 1994)

Answer 5

-We steer using geometrical properties of an optical feature of the bend (tangent point)
Land & Lee (1994):
-eye tracking while driving
-80% of time fixated on tangent point when entering bends

Question 6

What properties of the Tangent point aid steering

Answer 6

-On a constant curvature bend it stays in the same place
-Angle of TP indicates curvature of upcoming bend

Question 7

What are the assumptions of the Tangent Point Model

Answer 7

1). Theres an inside road edge
2). TP is not obscured
3). We ‘know’ distance from the edge
4). Bend is constant curvature
5). Steering directly related to TP angle

Question 8

Outline the Active Gaze Model

Answer 8

-Doesnt require road edges
-Simply fixate and track a point to steer through it
-Use perceptual inputs to change current steering response

Question 9

Outline Robertshaw & Wilkie (2008)

Answer 9

Instructed ppts to steer in middle or the road
-Exp1- free gaze mainly directed to middle of the road
-Exp2- When gaze forced onto TP, no improvement in steering performance

Question 10

Outline the evidence for the Active Gaze Model

Answer 10

-TP fixation in Land & Lee may be because ppts not instructed on trajectory
-AG model predicts TP fixation if wanting to take the ‘racing line’
-Land & Lee didnt record trajectories
-Wilkie et al. (2010)

Question 11

Outline Wilkie et al. (2010)

Answer 11

-Asked ppts to maintain lane position or take racing line
-Found gaze matched trajectory
-When asked to maintain different positions, gaze looked towards future path, not TP
:( - causality unclear (trajectory changes gaze or vice versa)

Question 12

Outline Readinger et al. (2002)

Answer 12

-Ppts asked to steer down straight road while looking at offset gaze angles
-Steered in direction of gaze, even when steering wheel mapping reversed

Question 13

Describe the findings from Robertshaw & Wilkie (2008) on fixation around bends

Answer 13

-Ppts steer towards fixation point in bends only on wide roads
-Visual info from road edges weaker on wide roads)

Question 14

Outline Kountouriotis et al. (2012)

Answer 14

-Used narrow road & manipulated road edge quality & fixation location
-Shifting gaze fixation away from centre increased steering bias, especially with lower quality
-Gaze direction had more influence with weaker road info
-Flexible weighting of info

Question 15

Outline Kountouriotis and colleague’s (2012) study dissociating gaze and steering

Answer 15

Exp1- free gaze (no fixation points)
-Had to maintain road position
-Manipulated road edge info quality, or removed an edge
-In some conditions, ppts didnt look where theyre going (with missing edge)
-But, made larger steering errors in these conditions
-Gaze and steering therefore linked