Study On Drivers’ Route Choice Behavior Under VMS Information

Traffic flow is the result of human behavior, which is closely linked to the choice and decision of travelers. The effects of route guidance system eventually depend on drives’ route choice behaviors under traffic guidance information. Thus, it is theoretically and practically useful to study drivers’ route choice behavior under traffic information. In the paper, the SP survey questionnaire and simulation experiment were conducted to study the behavior characteristic and decision-making mechanism, and the conclusions were further adopted to put forward some rational assumptions and cognitive process of route choice under the framework of cognitive psychology. The route choice model more in line with actual was proposed. Finally, multi-agent simulation under the microscopic simulation platform was used to optimize the traffic information release strategy. The main research work and the innovation were summarized as follows:1. A SP survey questionnaire was designed and conducted to collect data of drivers’ route choice behavior. Based on the definition of compliance rate, which was regarded as the dependent variable, an ordinal regression model was used to analyze the influencing factors of route choice behavior. The result showed age, income, character, occupation, driving experience, route choice behavior and the trust degree on VMS significantly influenced the effects of traffic information on drivers.2. Drivers’ risk attitudes in route choice behavior were analyzed based on the framework of prospect theory. Parameters for risk attitudes were estimated and verified through the empirical data from questionnaire survey and field spot. Moreover, the heterogeneities among drivers were considered. Drivers were classified into several typical groups according to their significant attributes, and their parameters of risk attitudes were estimated and verified separately. The result showed the heterogeneities among older drivers were more obvious, and drivers’ risk aversion to time gains decreased with the increasing of income.3. A route choice simulator was developed to carry out the simulation experiment of driver’s day-to-day route choice behavior under traffic information. Different from the data collected by questionnaire survey in different contexts, which was crosssection data, the data collected by experiment was dynamic and evolutionary, which covered the reactions to traffic information and learning from travel experiences and traffic information. The effects of traffic information, travel experience, the information accuracy and the habit on expected travel time updating and route choice were mainly studied. The result showed habit obviously affected drivers’ route choice behavior, and the effect increased with the increasing of the complexity of the decision environment.4. Benefit from the theory of cognitive psychology and SOAR cognitive architecture, the cognitive processes of route choice behavior, including perception, working memory, long-term memory, chunking, reinforcement learning and thinking etc., were described based on the conclusion and implication from the investigation and experiment.5. The multi-agent simulation model was validated by the comparison between the simulation result and the collected data. The result indicated that the route choice model within the framework of SOAR was effective in reflecting the actual route choice decision-making process. And the simulations were carried out under real-time traffic information and predicted information, respectively, the evolution of drivers’ route choice behavior and the traffic flow were compared to optimize the information release mode.