Driving behavior studies based on driving simulator experiments

Abstract

Driving behavior has been identified as a major contributing factor to critical accidents and road crashes. In Hong Kong, most road crashes (78.58%) involving traffic casualties are considered to be the drivers’ fault. A driving simulator approach was thus used to investigate driver behavior in distracted driving scenarios and scenarios involving signal countdown displays (SCD) at junctions.

Distracted driving is considered to be one of the most common causes of traffic crashes. Two distraction experiments of driving while distracted by texting messages and using hands-free mobile phones were conducted. In the experiment on reading and typing text messages, the drivers’ reaction time (RT), driving lane undulation (DLU), driving speed fluctuation (DSF), and car-following distance (CFD) between the test and leading cars were analyzed according to the following conditions: (1) no distraction, (2) reading and typing Chinese text messages and (3) reading and typing English text messages. The results indicated that reading and typing text messages while driving should be prohibited regardless of whether Chinese or English is used, as they cause remarkable impairments in driving performance. In the experiment on hands-free mobile phone use, the drivers’ RT, DSF, car-following distance undulation (CDU), and car-following time headway undulation (CTU) were used to measure and compare driving performance under the distraction conditions of no distraction (ND), normal conversation (NC), and seven levels of mathematical calculation tasks. The causal relationship between the calculation levels and driving performance factors were also investigated using a structural equation model (SEM). The study revealed that normal conversation on a hands-free mobile phone impaired driving performance to a degree equivalent to the distraction caused by level 3 mathematical calculations. In addition, the seven levels of calculation tasks can be applied to measure the distraction effects of other driving behaviors.

Road junctions and pedestrian crossings continue to be the locations where traffic accidents are most likely to occur. To better explore the effects of a SCD at a junction on driving performance, the driving behavior at green signal countdown displays (GSCD) was compared to that at conventional green light signal displays (GL), and the driving behavior at red signal countdown displays (RSCD) was compared to that at conventional red light signal displays (RL). Statistical tests of proportion were applied to compare the before-and-after installation effects. GSCD installation worsened the bimodal distribution of crossing speed at the stop line, in which the difference between the crossing speeds of the two sub-populations, risk-averse and risk-taking drivers, associated with the bimodal distribution was increased, and there was a tendency to encourage higher crossing speed for risk-taking drivers. Explanatory factors in terms of road condition, driving behavior, and individual characteristics were also included in the fixed-effects logistic regression models. GSCD installation encouraged stop decisions and reduced red-running violations, while RSCD installation encouraged deceleration to cross the junction instead of coming to a complete stop and increased the number of red-running violations.

This study was performed in an attempt to explore the effects of driving performance factors on driving behavior and to develop quantitative assessment models to evaluate driving behavior. The results of the distraction experiments were then applied to evaluate the distraction levels, and provide evidence for road safety legislation and education. The findings of the GSCD and RSCD driving simulations were used to assess the factors influencing driver behavior, and to evaluate the pros and cons of further SCD installation. Both findings are feasible and suitable for application in Hong Kong and overseas.