Study On Spatiotemporal Mining And Monitoring Method Design About Travel Behavior Base On Trajectory Data

Traveling is an important part of human activities and life style,and orderly and safe travel contributes to the harmony and stability of our society,and the happiness and security of human life.Travel behavior is a highly broad and multifaceted research subject that may differ in research objectives,methods,and application fields.However,the rapid developments and recent advances of the Internet,Internet of Things,mobile communications,geographic information systems,and the continuous improvement of big data mining,spatiotemporal analysis,distributed computing,high-performance computing and related technologies,have provided significantly enhanced data and methodological supports for the study of travel behavior.The present study takes the approach of examining the trajectory data generated by the residents' travel behavior and analyzing the complete process of travel between the origin and destination.From a geographical perspective,travel behavior is a process in which people interact with geographical time and space,so it follows the distribution law both in time and space.However,due to the differences in the purpose,mode and time of the travels,the intensity of spatiotemporal law in different travel behaviors can differ significantly.By combing the intensity and the relationship between the time and the spatial distribution laws,this paper consolidates three trajectory data-based travel behavior analysis models,namely time-dominant model,spatial-dominant model and spatiotemporal hybrid model.The applicability of the three perspective was discussed from both theoretical and methodological aspects.The problem requirements of different travel behaviors were discussed by combining with case studies.The design,optimization and combined use of the spatiotemporal analysis methods of different models were then tested.Finally,the developed model was used to analyze and mine the spatiotemporal laws underlined in the trajectory data,to regulate and intervene various travel behaviors,and to establish the application mode of the model.This work the developed model accelerates the management of travel behavior to a more intelligent,humanized and efficient level,and makes all kinds of travel to be safer,more reasonable and happier experiences.The research content of this paper includes three spatiotemporal analysis models and multiple travel modes.In the time-dominant model,the application of the time distribution law was divided into two types: the overall application and the segmentation application.In this model,by taking the random travel behavior associated with AOIs and the fixed-route travel behavior as examples,the social function identification of AOIs and the abnormal state detection of fixed-route vehicles were realized.In the spatial-dominant model,the single travel behavior of net carhailing and the mixture of community correction were used as examples,and by examining travels associated with different types of crime categories,the effects of geographic background zoning on travel behavior were examined and the background zoning difference of travel behavior was quantified,and both were used to evaluate the quality of travel behavior.The spatiotemporal hybrid model expounds on how to jointly realize the mining and supervision of travel behavior based on time law and space law.This was established by using the taxi travel as an example,focusing on the regulation of travel behavior and the monitoring of its access status at origin and destination.The main innovation and work of this study are as follows:1)The spatiotemporal analysis model of travel behavior.A spatiotemporal analysis model framework of the travel behavior process based on the trajectory is proposed from the perspective of geography.According to the intensity and the degree of correlation between the temporal and the spatial distributions,the models are divided into three types: time-dominant model,spatial-dominant model and spatiotemporal hybrid model.The conceptual model,the implementation process and the application cases of the three types of the spatiotemporal analysis model were discussed.2)Travel behavior analysis dominated by time distribution law.Based on the timedominant model,a method of AOI social function type identification based on k NN and spectrum matching is herein proposed.In this method,the travel behavior track associated with a typical AOI is trained with the time spectrum of pick-up and drop-off points to form the standard spectrum of various AOI.Continuous matching of the similarity between the spectrum of drop-off points cluster calculated by clustering and the spectrum itself results in the dynamic identification of the associated AOI types.This is a novel monitoring method of fixed-route travel behavior in real time based on the time distribution law,by which the controllability of local time distribution can be obtained.3)Travel behavior analysis dominated by spatial distribution law.Based on the spatial-dominant model,a method of travel behavior evaluation based on geographic background zoning is proposed.Based on the geographic area of the travel behavior track,the method classifies the level and the category of the geographical regional difference characteristics,and sets them as the qualitative evaluation standard of travel behavior for the execution of the management of travel behavior.The feasibility of this analysis model was verified by using the single and hybrid travel mode as examples for the state evaluation of travel behavior.4)Analysis of travel behavior under the action of spatial and time distribution.This is a new method of intermediate process scheduling of travel behavior and dynamic monitoring of traffic state at origin and destination,based on the spatiotemporal hybrid model.Under the action of spatial and time distribution,the dynamic scheduling scheme divides the time into several parts for management based on the time distribution law.Meanwhile,the dynamic scheduling scheme searches hot and cold points based on the spatial distribution law to realize the regional and the macro-control of the travel behavior.The access status monitoring method is realized by accurate extraction of entrance and exit of AOI according to the spatial distribution law,and by the accurate monitoring of access status based on the time distribution law.The present study not only provides a new spatiotemporal analysis method but also optimizes and combines the existing spatiotemporal analysis methods.This novel approach can be applied to several typical travel behaviors in different industries.A series of supervision and scheduling methods are also designed to help realize abnormal state detection,security assurance and dynamic management in the travel process,based on their respective characteristics.