Analysing driver comprehension of formats for presenting traffic information on variable message signs : towards stated preference approach for travel behaviour research

Abstract

Providing drivers with real-time dynamic traffic information by Variable Message Signs (VMSs) has been developed as an important initiative in traffic management of highways. The success of a VMS system highly depends on the quality of the messages as drivers' route choice behaviour will be affected by the available traffic information. Some messages do have much more influence than others, especially when they are presented in a particular format. To produce optimal operation, appropriate formats for presenting traffic information should be carefully selected to adopt users' need. The effectiveness of the system will be reduced if users misinterpret the contents of the messages. However, much attention has been given in the literature to forecast the use or penetration of VMSs, while proportionate attention has not been given to predict how drivers will respond to information presented in different formats. Analysis of driver route choice behaviour is a difficult task and the condition will be much more complicated if traffic information is also provided to drivers. The level of inter-relationship between the information in different formats cannot be measured explicitly by conventional methods. This study attempted to examine driver comprehension and en-route choice behaviour in response to traffic information that was presented in three formats, namely; Numerical format (presenting travel time in a number), Words format (presenting congestion level by description) and Signal format (indicating congestion level by number of switch-on-signals). A stated preference (SP) survey with well-designed attributes regarding traffic conditions and comprising a total of 475 successful cases was conducted to measure drivers' utilisation for travelling on two designated routes. SP techniques provide data on the underlying priorities that determining drivers' behavioural intentions. Discrete choice analysis was employed to develop binary logit models with respect to drivers' preference in different circumstances. Each driver was expected to choose the route that generated the highest utility to he or she. If this was not the case, the choice could be considered irrational. When a driver selected the route, which he or she had the highest utility, his or her behaviour could be quantified and expressed with a utility function. The overall utility of drivers for travelling on the routes could therefore be affected by their comprehension of the message formats. The vector-valued variables employed in the utility functions were addressed by well-defined attributes. The explanatory variables employed in the models including those refer to respondents' socio-economic background and those characterising the transport network. To demonstrate the explanatory power of the models, respondents were also grouped into different categories regarding their familiarity with the road network and listening propensity to radio traffic news. It was understood that drivers' perception process affected their utilisation of the travel conditions of the routes. Drivers had different trade-off attitudes to the time and costs related attributes when they were influenced by identical traffic information which was presented in different formats. It was concluded that drivers had different comprehension and perception mechanism to the formats for presenting traffic information on VMS system. The message formats could induce certain bias towards a route in one's decision-making process.