Mining And Analysis Of Spatio-temporal Patterns Of Urban Human Activities Based On Floating Car Data

Human movement pattern is a collection of temporal and spatial characteristics of travel behavior,which not only reflects the dynamic rhythm of urban crowd travel in time and space,but also reflects the interaction between people and urban space.With the accelerating process of urbanization,people’s movement behavior in the city has become diverse and complex,which is a big challenge to the balance between urban spatial structure and people’s demand,and has resulted in a series of problems of interaction between people and the city,such as urban heat island,traffic congestion,unbalanced resource allocation and so on.The mismatch between the pattern of crowd movement and urban spatial structure is an important reason for a series of urban problems.It is of great practical significance to deeply understand the mode of urban crowd movement for guiding the development of urban traffic,planning and emergency plans.The development of information technology makes it possible to collect largescale spatiotemporal trajectory data(such as mobile phone signaling data,check-in data,taxi trajectory data,etc.).Compared with traditional survey data,the characteristics of large data sample size and strong real-time nature help researchers to perceive and understand the dynamic changes of urban population and urban spatial structure from a finer temporal and spatial scale.Vehicle GPS taxi trajectory data,as a part of the intelligent transportation system,contains vehicle load information and location information,and has gradually become an important data source for people to explore the urban spatial structure and reveal the laws of human activities.However,the existing research mainly focuses on the temporal and spatial heterogeneity of the movement trajectory,and the research on the movement patterns of different actors is not thorough enough,and it lacks an effective combination of relevant semantic information such as the description of the urban environment.For different groups of people in the city,how to quickly and effectively extract meaningful information from the massive trajectory data,deepen the understanding and understanding of the behavior patterns of specific groups of people,dig out deep-level spatial information and serve related applications,is an important aspect of current human travel research.Consequently,based on the massive floating car data,combined with the basic geographic information data of other cities,this paper focuses on exploring the movement patterns of two kinds of actors(drivers and residents)and the interaction characteristics with the urban spatial structure.The main research work includes:1.Research on similarity measurement mining method of trajectory data.Aiming at the characteristics of taxi trajectory data,a similarity measurement algorithm suitable for taxi trajectory data is proposed to identify the deep semantic information contained in the trajectory data.First,based on the spatial heterogeneity of the urban road network distribution,the thesis divides the study area into different subregions,the study area is divided into different sub regions,and sets the grid size of the sub-regions according to the distance between the road networks.Secondly,comprehensively considering the spatial position and self-intersection characteristics of the trajectory,a set of algorithms for measuring the similarity of the spatial trajectory is proposed.And use other algorithms to cross-validate the effectiveness of the algorithm.2.Analysis of the movement pattern of the driver group.Based on the massive taxi trajectory data,the operation characteristics and movement patterns of drivers are mined and analyzed.First,analyze the characteristics of taxi operating time and mileage.Secondly,according to the moving parameters of taxi trajectory(such as speed,distance and time),the method of vehicle stop recognition is optimized.The stability of vehicle stop is analyzed by setting speed threshold and noise threshold.The travel trajectory is segmented and the data set of stop trajectory is constructed.On this basis,the spatial analysis method is used to mine the spatial and temporal characteristics of driver group stop behavior.Finally,on the basis of identifying the stopping points of drivers,the spatial distribution of taxi drivers in Shanghai is analyzed from the perspective of driver community,and the spatialtemporal similarity of community vehicle operation mode is verified by setting up comparative experiments.3.Research on the movement pattern and space interaction of urban residents.Based on the taxi trajectory data in June 2018,this paper analyzes the movement patterns and spatial interaction characteristics of urban residents.Firstly,according to the data of taxi trajectory and the position of pick-up and drop-off(OD matrix),the spatial and temporal statistical characteristics of residents’ trip volume,distance and time are analyzed;secondly,from different spatial scales,the spatial statistics and nuclear density analysis methods are adopted to analyze the spatial distribution characteristics of r residents’ movement,and identifying urban hot spots and road sections.Third,combine the purpose of urban residents’ activities with urban POI data to identify the urban functional structure and verify the effectiveness of the method.Finally,combining urban administrative divisions and urban functional structure data,the spatial interaction characteristics of urban residents’ movement are analyzed from different time and space scales.4.Case study on the mining of emergency travel pattern of urban residents.Taking urban medical resources as an example,using floating car trajectory data to perceive travel in urban hospitals.By exploring the spatial interaction between residents and urban hospitals,the medical behavior of the residents and the characteristics of urban medical resource allocation are analyzed.First of all,a method is proposed to identify the emergency medical trip at night and where the patient comes from.Then,the alpha-shape(boundary extraction algorithm)and Isolation Forest algorithm are used to divide the hospital service area,and further discuss the residents’ medical behavior choice preference.Finally,the spatial accessibility of medical institutions in Shanghai was calculated and evaluated.