Dynamic changes of traffic features in unstructured road networks challenge the scene-cognitive abilities of drivers,which brings various heterogeneous traffic behaviors.Modeling traffic with these heterogeneous behav...Dynamic changes of traffic features in unstructured road networks challenge the scene-cognitive abilities of drivers,which brings various heterogeneous traffic behaviors.Modeling traffic with these heterogeneous behaviors would have significant impact on realistic traffic simulation.Most existing traffic methods generate traffic behaviors by adjust-ing parameters and cannot describe those heterogeneous traffic flows in detail.In this paper,a cognition-driven traffic-simulation method inspired by the theory of cognitive psychology is introduced.We first present a visual-filtering model and a perceptual-information fusion model to describe drivers'heterogeneous cognitive processes.Then,logistic regression is used to model drivers'heuristic decision-making processes based on the above cognitive results.Lastly,we apply the high-level cognitive decision-making results to low-level traffic simulation.The experimental results show that our method can provide realistic simulations for the traffic with those heterogeneous behaviors in unstructured road networks and has nearly the same efficiency as that of existing methods.展开更多
基金supported by the National Natural Science Foundation of China under Grant Nos.61602425,61572445 and 61822701the Doctor Fund of Zhengzhou University of Light Industry of China under Grant No.2015BSJJ007。
文摘Dynamic changes of traffic features in unstructured road networks challenge the scene-cognitive abilities of drivers,which brings various heterogeneous traffic behaviors.Modeling traffic with these heterogeneous behaviors would have significant impact on realistic traffic simulation.Most existing traffic methods generate traffic behaviors by adjust-ing parameters and cannot describe those heterogeneous traffic flows in detail.In this paper,a cognition-driven traffic-simulation method inspired by the theory of cognitive psychology is introduced.We first present a visual-filtering model and a perceptual-information fusion model to describe drivers'heterogeneous cognitive processes.Then,logistic regression is used to model drivers'heuristic decision-making processes based on the above cognitive results.Lastly,we apply the high-level cognitive decision-making results to low-level traffic simulation.The experimental results show that our method can provide realistic simulations for the traffic with those heterogeneous behaviors in unstructured road networks and has nearly the same efficiency as that of existing methods.
