This is the author s version of a work that was submitted/accepted for publication in the following source: Haque, Md. Mazharul, Washington, Simon, Ohlhauser, Amanda, & Boyle, Linda (2012) Effects of Mobile Phone Distraction at the Onset of Amber Light: Analysis of Driving Simulator Data. In QUT Transport Policy, Planning and Engineering Symposium, 15 Jun 2012, Brisbane. (Unpublished) This file was downloaded from: https://eprints.qut.edu.au/51261/ c Copyright 2012 the authors. Notice: Changes introduced as a result of publishing processes such as copy-editing and formatting may not be reflected in this document. For a definitive version of this work, please refer to the published source:
Effects of Mobile Phone Distraction at the Onset of Amber Light: Analysis of Driving Simulator Data Dr Md. Mazharul Haque, Prof. Simon Washington, Amanda aven & Prof Linda Boyle MD. MAZHAUL HAQUE, PhD Civil Engineering and the Built Environment & Centre for Accident esearch and oad Safety - Queensland (CAS-Q) Queensland University of Technology (QUT) Queensland University of Technology Background AGENDA Experimental Setup Statistical Modeling esults & Discussions Conclusion 1
BACKGOUND U.S. roadways Driving distraction claimed about 5,474 fatalities and 448,000 injuries in 2009. Cell phone distraction alone caused about 995 people to be killed and 24,000 to be injured. Australian roadways About 2,400 distraction related incidents were reported in NSW (2002) mostly by young drivers using cell phone. In 2010, ACT police issued 2,323 infringement notices and 445 caution notices to drivers using cell phone. Media Watch BACKGOUND The Daily Telegraph June 01, 2012 The Courier Mail November 11, 2011 2
BACKGOUND esearch on Cell Phone Distraction Speed Control Vehicle Control Lane Departures Braking Performances eaction Time Behavior at Signalized Intersections Stop/Go Decision at Amber Light An improper decision i at amber light may cause red light running or abrupt stop. Amber lights were problematic for both younger and older drivers (Koneci et al., 1976; Edwards et al., 2003). BACKGOUND Stop/Go Decision at Amber Light Novice drivers were more likely to run through an amber light (Senserrick et al., 2007). Higher time to stop line decreased the likelihood of running through amber light (Caird et al., 2007). Cell phone distraction increased the likelihood of failure to detect traffic signals (Strayer and Johnston, 2001) esearch Question In general, phone conversation increases perception response time to critical driving events (e.g., Consiglio et al., 2003). How do cell phone distracted drivers response at the onset of amber light? 3
BACKGOUND OBJECTIVE To examine the decisions of distracted drivers at the onset of amber light. SCOPE A group of distracted drivers were exposed on a series of signalized intersections using the National Advanced Driving Simulator of IOWA. EXPEIMENTAL SETUP Driving Simulator (NADS 1) 360 0 driver view in a 24 feet dome Dome can rotate up to 330 0 in its vertical axis 13 degree of freedom motion base Two belts driven beams to provide lateral and longitudinal accelerations Data are recorded at rates up to 240 Hz 4
Participants EXPEIMENTAL SETUP Four age groups: Novice, 16-17 years old; Males = 11, Females = 0 Licensed between 4 and 8 weeks Younger, 18-25 years old; Males = 9, Females = 8 Middle, 30-45 years old; Males = 9, Females = 8 Older, 50-60 years old; Males = 8, Females = 5 Procedure Briefing about how to use the simulator and cell phone apparatus A familiarization drive before the experimental drive. Procedure EXPEIMENTAL SETUP Six signalized intersections on the simulated driving route 2 in the baseline condition (with no device) 4 in the handheld phone condition Traffic lights changed from green to amber when the driven car was 3.0 seconds from the stop line (3 intersections) 3.75 seconds from the stop line (3 intersections) Cell Phone Conversation Cognitive type of conversation task Drivers needed to determine whether a sentence made sense or not. 5
Variables STATISTICAL MODELING Stop/Go decision (Binary response) Phone condition (baseline, handheld) Driver s Age group (novice, younger, middle, older) Gender (male, female) Time to stop line (TSL): varies between 2.46 and 3.85 seconds Statistical Model Driver s decision = f (age, gender, TSL, phone condition) epeated measures Logistic egression model with a combination of a Decision Tree analysis. Decision Tree ESULTS Interaction Variable1: A novice driver without phone conversation and distance from the stop line at yellow light 188.5 ft 6
Logistic egression ESULTS (CNTD.) Time to Stop Line (TSL) 1 sec increase in TSL was associated with 9.7 fold increase in stopping at amber light. Handheld (HH) Phone Condition Overall the phone conversation decreased the likelihood of proceeding through amber light. TSL x HH 1 sec increase in TSL was associated with 10.6 times increase in running through an amber light for distracted drivers. Driver s Age The middle aged group was the least likely to run an amber light, followed by older drivers, and younger drivers. Logistic egression ESULTS (CNTD.) Age x TSL The likelihood of running through amber light increased with time to stop line for younger and older drivers Higher Order Interaction Variables A novice driver was more likely to run through amber when the distance was 188.4 ft. A young driver was about 11 times more likely to run the amber light when the speed was > 44.9 mph. An older male driver was more likely to run through the amber light when the speed was > 46 mph. TSL: Time to Stop Line 7
Logistic egression ESULTS (CNTD.) Gender Female drivers were more likely to run through an amber light. Age x HH Novice and young drivers were more likely to run the amber light when distracted by cell phone conversation. Age x HH x TSL As TSL increases, distracted novice and young drivers were less likely to proceed through the amber light. TSL : Time to Stop Line HH : Handheld Phone DISCUSSION 1 Non Distracted eference Group Distracted eference Group Non Distracted Male Novice Distracted Male Novice Non Distracted Male Younger Distracted Male Younger Non Distracted Male Older Distracted Male Older Non Distracted Female Younger Distracted Female Younger Non Distracted Female Older Distracted Female Older nning through an amber lig ght Probability of ru 0.8 0.6 0.4 0.2 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 Time to the stop line (sec) 8
1 DISCUSSION (CNTD.) Non Distracted eference Group Non Distracted Male Novice Distracted eference Group Distracted Male Novice ty of running through an amber light Probabilit 0.8 0.6 0.4 0.2 70% 30% 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 Time to the stop line (sec) ity of running through an amber light Probabili 1 0.8 0.6 0.4 0.2 48% DISCUSSION (CNTD.) Non Distracted Male Younger Non Distracted Female Younger Distracted Male Younger Distracted Female Younger Avg. eduction 30% 62% 40% 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 Time to the stop line (sec) 10% 9
DISCUSSION (CNTD.) lity of running through an amber light 1 0.8 0.6 0.4 0.2 Distracted Male Older Distracted Female Older Probabi 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 Time to the stop line (sec) CONCLUSION Drivers responses to distraction and running an amber signal vary considerably across age and gender. Young and novice drivers are more likely to run through an amber light while distracted by cell phone conversations. Overall, drivers tend to compensate for the increased risk of phone conversation by running the amber light less often. isk compensation is also exhibited across most tested groups. Whether the risk compensation is sufficient to offset crash risks we don tknow; but on road situation ti suggest it is not. The behavior of older drivers and gender differences require further investigations. 10