Temporal patterns of driving fatigue and driving performance among male taxi drivers in Hong Kong: A driving simulator approach
Introduction
As one of the most important modes of public transport, taxis play a key role in the modern transportation system by offering passengers flexible, comfortable, point-to-point travel service (Wu et al., 2016). As the global taxi industry’s revenues have grown, serious safety concerns regarding taxi trips have been raised (Baker et al., 1976; Meng et al., 2017b). According to the Transport Department of Hong Kong (2016), 3928 crashes involving taxis occurred in Hong Kong in 2016, resulting in 5352 casualties in the taxis involved. Both figures rank second among the 17 classes of vehicles, trailing only private cars. From 2007–2016, the number of crashes involving taxis in Hong Kong rose by 18.3%. Although the efficiency and the comfort level of trips were enhanced by improvement of taxi services, the frequent taxi crashes and the large number of casualties still puzzle transport managers in Hong Kong and worldwide (Meng et al., 2017b).
Taxi drivers’ aggressive driving attitudes and risky driving performance have apparently led to an increase in hidden crash risk and have been frequently investigated (Machin and De Souza, 2004; Rosenbloom and Shahar, 2007; Shams et al., 2011; Cheng et al., 2016). Rosenbloom and Shahar (2007) studied the attitudes toward traffic violation penalties between male taxi drivers and nonprofessional drivers in Israel and thus measured their legal obedience levels. The results of a survey with 80 participants showed that taxi drivers judged the penalties as less severe than nonprofessional drivers, especially those with penalty conditions of low and medium severity, possibly as a result of different driving attitudes: taxi drivers may be willing to risk violating traffic rules to increase their profits. This hypothesis was verified in a more recent study by Cheng et al. (2016), in which impulsivity and risky decision-making tendencies were compared in 30 taxi drivers, including 15 traffic offenders and 15 non-offenders. The taxi drivers with traffic offence records were found to be less sensitive to the consequences of risky behavior and were more profit-driven than their non-offending counterparts. These findings not only unveiled the possible causes of taxi drivers’ aggressive attitudes as hypothesized by Rosenbloom and Shahar (2007) but also further proved that the profit-making nature of taxi services resulted in taxis drivers’ risky decision-making and driving performance. To more specifically investigate taxi drivers’ driving performance, Wu et al. (2016) conducted a driving simulator study with two simulated scenarios: red-light running violation and crash avoidance at intersections. Taxi drivers ran red lights with a significantly greater frequency than non-professional drivers, indicating that taxi drivers were more inclined to cross the intersection during amber light and thus displayed more violating behaviors; however, taxis drivers showed better crash avoidance behavior at the simulated intersections.
It has long been argued that fatigued driving may lead to risky driving performance and aggressive driving attitudes because driving fatigue can reduce a driver’s alertness and cause poor psychometric conditions (Dalziel and Job, 1997; Merat and Jamson, 2013; Wu et al., 2016). Indeed, professional drivers such as taxi drivers commonly drive at a high fatigue level because they tend to drive continuously for long hours with a high working intensity because of the profit-driven nature of their driving. Dalziel and Job (1997) examined the relationships between fatigue-related variables and traffic crash involvement in a survey of 42 taxi drivers in Sydney, Australia, and concluded that longer driving hours produced higher crash risks and that taking longer breaks during a shift could help alleviate the situation. Similarly, prolonged driving hours were found to contribute to driving fatigue among taxi drivers by Meng et al. (2015) based on a survey in which taxi drivers’ fatigue perception was compared with that of truck drivers. The researchers also found that taxi drivers reported significantly more fatigued driving experiences and greater crash involvement rates than truck drivers. In addition to the fatigue gained through driving, disordered night-time sleep was also found to contribute to drivers’ daytime driving fatigue (May et al., 2016). Firestone et al. (2009) surveyed 241 taxi drivers in Wellington, New Zealand, and showed that obstructive sleep apnea syndrome was prevalent among taxi drivers, especially among the Maori and Pacific ethnicities.
Although it seems plausible that longer driving hours may cause greater driving fatigue in taxi drivers, the pattern of driving fatigue and driving performance along with driving hours in a working shift has never been investigated. The origins of driving fatigue have been shown to be comprehensive (Meng et al., 2015), and continuous long-hour driving is not its only cause. Sleep disorders, taking breaks during driving, driving intensity, and self-perceived fatigue can all affect drivers’ fatigue levels and fatigued driving performance (Ting et al., 2008; Merat and Jamson, 2013; Huffmyer et al., 2016; May et al., 2016). Moreover, taxi services in Hong Kong are rather flexible, so each driver can take a break whenever he feels fatigued and may thus seek his own balance between making profits and maintaining alertness and driving safety. Therefore, taxi drivers’ fatigue levels and driving performance over time during a shift remain subtle if not quantitatively modeled.
According to Transport Department of Hong Kong, approximately 15% of taxi drivers in Hong Kong are female (6000 of 40,000 valid taxi driver licenses), but a large majority are part-time drivers who drive taxis infrequently. Moreover, according to the road traffic crash records of the Hong Kong Police Force, 4163 taxis were involved in road traffic crashes in 2011, of which 98.2% (4088 taxis) were driven by a male driver when the crash occurred. Therefore, considering the low percentage of female taxi drivers and the much higher rate of crash involvement of male taxi drivers in Hong Kong, this study focused on male taxi drivers only. In this paper, a driving simulator experiment and a fatigue survey were conducted among male taxi drivers to identify the role of driving hours in taxi drivers’ fatigue levels and driving performance. A following-braking scenario was applied, and the drivers’ driving and reaction behaviors were recorded and analyzed. The Brief Fatigue Inventory (BFI) was used to evaluate the drivers’ fatigue levels. Each taxi driver was required to participate at three points: before, during, and after a normal work shift, to account for the effects of driving/working hours on their driving performance and fatigue levels. Policy implications were proposed based on the results of the analyses to cope efficiently with the taxi drivers’ driving fatigue and further alleviate the taxi safety situation in Hong Kong.
Section snippets
Participants
Fifty male taxi drivers between 23 and 66 years of age (mean 45 years) were recruited in Hong Kong. All recruited drivers were legal Hong Kong residents with a valid taxi driving license issued by the Transport Department of Hong Kong. Each driver was asked to visit the laboratory three times: before, during, and after their normal working shifts. All drivers were asked to refrain from consuming caffeinated drinks and alcohol during the 24 h before their scheduled experiments. Free parking
Modeling driving performance with driving fatigue
To investigate the effect of the taxi drivers’ fatigue level on their driving performance with unobserved heterogeneities, various measures dependent upon driving performance recorded in the driving simulator experiments were modeled with the participants’ self-reported fatigue at the time of the experiment and their demographic factors using RE models. A dummy independent variable, serious fatigue, was adopted to represent the driver’s level of fatigue. The value of this variable was 1 if the
Effect of driving fatigue on driving performance
Based on the results shown in Table 2, male taxi drivers’ self-reported fatigue levels were significant when modeling BRT_50, SDLane_50, SDLane_80, and VarSteer_80, whereas other confounding demographic variables were incorporated and unobserved heterogeneities were considered. The coefficients of serious fatigue were all significantly positive in the four listed models, which means that in general, the more seriously fatigued drivers tended to have worse driving performance than the drivers
Conclusions
In this study, a questionnaire and a driving simulator experiment with FB scenarios were designed to define the temporal patterns of the fatigue levels and driving performance of 50 male Hong Kong taxi drivers in their working shifts. The same measurements were conducted for each participant at three times: before, during, and after his working shift. The questionnaire recorded the drivers’ demographic information, such as age, daily net income, and daily sleeping hours, and their BFI scores at
Acknowledgments
This study was supported by grants from the University Research Committee of The University of Hong Kong (201511159015); National Natural Science Foundation of China (71561167001); Research Grants Council of Hong Kong (N_HKU707/15); National Natural Science Foundation of China (71671100). The first and second authors were also supported by the Research Postgraduate Studentship and by the Francis S Y Bong Professorship in Engineering, respectively. We express special gratitude to the Hong Kong
References (28)
- et al.
Driver performance in the moments surrounding a microsleep
Transp. Res. Part F Traffic Psychol. Behav.
(2008) - et al.
Impulsivity and risky decision making among taxi drivers in Hong Kong: an event-related potential study
Accid. Anal. Prev.
(2016) - et al.
Motor vehicle accidents, fatigue and optimism bias in taxi drivers
Accid. Anal. Prev.
(1997) - et al.
Obstructive sleep apnoea among professional taxi drivers: A pilot study
Accid. Anal. Prev.
(2009) - et al.
Are exercise stress tests appropriate for assessing myocardial ischemia in patients with major depressive disorder?
Am. Heart J.
(2004) - et al.
A simulation study of the effects of alcohol on driving performance in a Chinese population
Accid. Anal. Prev.
(2016) - et al.
Predicting health outcomes and safety behaviour in taxi drivers
Transp. Res. Part F Traffic Psychol. Behav.
(2004) - et al.
The deterioration of driving performance over time in drivers with untreated sleep apnea
Accid. Anal. Prev.
(2016) - et al.
Driving fatigue in professional drivers: a survey of truck and taxi drivers
Traffic Inj. Prev.
(2015) - et al.
Gas dynamic analogous exposure approach to interaction intensity in multiple-vehicle crash analysis: case study of crashes involving taxis
Anal. Methods Accid. Res.
(2017)
Occupant-level injury severity analyses for taxis in Hong Kong: a Bayesian space-time logistic model
Accid. Anal. Prev.
The effect of three low-cost engineering treatments on driver fatigue: a driving simulator study
Accid. Anal. Prev.
Differences between taxi and nonprofessional male drivers in attitudes towards traffic-violation penalties
Transp. Res. Part F Traffic Psychol. Behav.
Taxi drivers’ views on risky driving behavior in Tehran: a qualitative study using a social marketing approach
Accid. Anal. Prev.
Cited by (52)
Driver fatigue in taxi, ride-hailing, and ridesharing services: a systematic review
2024, Transport ReviewsJointly analyzing freeway primary and secondary crash severity using a copula-based approach
2023, Accident Analysis and PreventionDistractions by work-related activities: The impact of ride-hailing app and radio system on male taxi drivers
2022, Accident Analysis and PreventionAn assessment of the effect of green signal countdown timers on drivers’ behavior and on road safety at intersections, based on driving simulator experiments and naturalistic observation studies
2022, Journal of Safety ResearchCitation Excerpt :For example, Islam et al. (2017) examined the driving behaviors of 55 drivers in the United States in a simulator experiment of multiple signal intersections with GSCTs, and concluded that GSCTs increased intersection safety by increasing drivers’ propensity to stop and decreasing their average deceleration rates. In comparison with naturalistic observation studies, driving simulator–based experimental studies allow the capture of detailed information on each driver’s driving behavior under specific circumstances, such as signalized intersections with GSCTs (Lee et al., 2002; Yan et al., 2007; Yan et al., 2018; Meng et al., 2019). Overall, the driving simulator–based experiments and naturalistic observation are complementary.
An improved automated braking system for rear-end collisions: A study based on a driving simulator experiment
2022, Journal of Safety Research