Determinants of driver response to variable message sign information in Athens

Abstract

Variable message signs -which comprise a type of advanced traveller information systems -can affect driver behaviour, especially considering route choice. Hence, their operation is integrated in traffic management strategies for the mitigation of traffic congestion. This research explores the factors determining driver response to variable message signs (VMSs) in the city of Athens. A stated preference questionnaire survey is undertaken and discrete choice analysis is performed towards this aim. More specifically, a random-effect ordered probit model is estimated that provides insight on the contributory factors that influence driver propensity to divert, when provided with information on incident occurrence via VMSs. Message characteristics, that is, incident type, impact and suggestion for an alternative route, trip characteristics, such as vehicle type, as well as, driver characteristics, such as driver age and income, have been found to affect driver behaviour. Furthermore, appropriate models are also estimated for subsets of the driving population (considering gender and age) and specific similarities and differences between the population behaviours are identified.

Analyzing drivers’ preferences and choices for the content and format of variable message signs (VMS)

“…(Wardman et al, 1997;Al-Deek et al, 2012). Delays attribute to accidents have the larger impacts on route choice than other stated causes, such as roadworks and congestion (Wardman et al, 1997;Chatterjee et al, 2002;Spyropoulou and Antoniou, 2014;Sharples et al, 2016). Jindahra and Choocharukul (2013) employed the mixed logit model to analyze drivers' preference for VMS contents, the results showed that drivers would like to receive the suggested route diversion information, qualitative delay information and road construction information when they are at congested conditions.…”

“…This includes: (i) revealed preference (RP) questionnaire surveys (Chatterjee et al, 2002;Kilpeläinen and Summala. 2007;Wang et al, 2009;Hassan and Abdel-Aty, 2011), stated preference (SP) questionnaire survey (Benson, 1996;Emmerink et al, 1996;Wardman et al, 1997;Al-Madani andAl-Janahi, 2002a, 2002b;Hassan and Abdel-Aty, 2011;Gan and Ye, 2011;Jindahra and Choocharukul, 2013;Spyropoulou and Antoniou, 2014;Ronchi et al, 2016) and a combination of SP and RP survey (Al-Deek et al, 2012), (ii) driving simulators (Mahmassani and Chen, 1991;Wang and Cao, 2005;Lai, 2008Lai, , 2012Shinar and Vogelzang, 2013;Williams et al, 2015;Sharples, et al, 2016;Roca et al, 2018 ), and (iii) the combination of questionnaire survey and driving simulators (Yang et al, 2005;Wang et al, 2007;Lai, 2010;Cristea and Delhomme, 2014;Zhong et al, 2016).…”

Powered‐two‐wheelers and smart cities: the case of variable message signs

“…Information provision on alternative routes is yet another significant parameter affecting driver diversion behaviour. Findings from different studies are in agreement, and suggest that drivers exhibit higher diversion propensity when such information is transmitted [26][27][28][29]31]. In a much earlier study, though, Benson [32] noted that motorists were not interested in the provision of alternative route information.…”

“…Poulopoulou et al [28] and Poulopoulou and Spyropoulou [29], on the other hand, found that incident type does not affect professional driver behaviour. Impact magnitude comprises yet another contributory parameter of driver diversion behaviour, regardless of the driver category explored [25][26][27][28][29][30]. Examples of impact magnitude include road or lane closure, travel times or delay being expressed either qualitatively (e.g.…”

“…In particular, the cause of congestion (also referred to as incident type) plays a significant role in driver diversion, with drivers exhibiting higher propensity to divert to specific incident incurred delay [23, 24]. Several different incident types have been explored including demonstration, accident and roadworks, most of which exhibit differences in drivers’ responses [25–27]. Poulopoulou et al [28] and Poulopoulou and Spyropoulou [29], on the other hand, found that incident type does not affect professional driver behaviour.…”