TRAFFIC SIMULATOR FOR EVALUATION OF TRAFFIC ACCIDENT INFORMATION SYSTEMS
Hiroyuki Mori, H. Tsuji, Hisato OHNO, Kazutoshi KAMIYA, K Goto
Abstract
Hiroyuki Mori, H. Tsuji, Hisato OHNO, Kazutoshi KAMIYA, K Goto
Abstract
It is important to evaluate the effectiveness of traffic accident information systems before the actual implementation of the systems; however, it is impossible to examine the systems in actual situations because of the risk of true crashes. Therefore, a traffic simulator is developed in which it is possible to: 1) cause traffic accidents under the situation of traffic flow in which plural cars run, and 2) simulate actual driver behaviors and vehicle movements of each car. Vehicle behaviors were observed in actual traffic flow, relative speeds and distances were measured between the following car and the leading car by the vehicle following experiment, and driver's consciousness concerning driving psychology was investigated. On the basis of the analysis with the observation results, a traffic flow generation model and several driver models, which describe driver behaviors such as car following and lane changing, are made and the parameters of the models are calibrated. Consequently, it is confirmed that the simulator has the ability to represent the traffic flow which agrees well with the actual ones. As an example, a simulation experiment of multiple rear-end collisions on an expressway is performed using the simulator, and it is shown that the simulator can evaluate the effectiveness of the traffic accident information systems.
A significance statement is not available in the OpenAlex record.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
It is important to evaluate the effectiveness of traffic accident information systems before the actual implementation of the systems; however, it is impossible to examine the systems in actual situations because of the risk of true crashes. Therefore, a traffic simulator is developed in which it is possible to: 1) cause traffic accidents under the situation of traffic flow in which plural cars run, and 2) simulate actual driver behaviors and vehicle movements of each car. Vehicle behaviors were observed in actual traffic flow, relative speeds and distances were measured between the following car and the leading car by the vehicle following experiment, and driver's consciousness concerning driving psychology was investigated. On the basis of the analysis with the observation results, a traffic flow generation model and several driver models, which describe driver behaviors such as car following and lane changing, are made and the parameters of the models are calibrated. Consequently, it is confirmed that the simulator has the ability to represent the traffic flow which agrees well with the actual ones. As an example, a simulation experiment of multiple rear-end collisions on an expressway is performed using the simulator, and it is shown that the simulator can evaluate the effectiveness of the traffic accident information systems.
Key concepts: Simulation, Driving simulator, Traffic flow (computer networking), Computer science, Traffic simulation, Traffic accident, Transport engineering, Engineering