Study On The Speed Periodicity Of Urban Road Network Based On Taxi Trajectory Data

Taxi trajectory data is a typical temporal and spatial sequence data, which largely reflects the behavioral characteristics of urban residents take a taxi in time and space.It will provide a useful reference for the development of urban traffic to carry on the in-depth mining and the analysis to the massive taxi trajectory data. At present, the research and application of taxi trajectory data has become a hot topic in related fields.Space-time data mining is using various data analysis tools and methods to mine information and laws implied in spatiotemporal databases.This paper takes Wuhan taxi trajectory as source data, extracts the road network,and the taxi record of the track data to determine the direction of travel. Based on the grid will quickly match the track points to the corresponding sections and then calculate the speed of the road at different times, to build the road speed signal. The road speed signal is converted from the time domain to the frequency domain by Fourier transform,periodic analysis using periodogram method, extract the road speed change cycle,mining the urban road traffic information. In general, this paper presents a set of urban traffic flow velocity change cycle model. The main research work is as follows:(1) Road centerline extraction based on taxi trajectory dataExtraction of urban road network by taxi trajectory data, construction of Delaunay triangulation based on taxi trajectory data. The triangular long edge threshold of the deleted triangulation is obtained by statistical analysis experiment, extracted the polygon of the road area, and simplified the processing of polygon edge roughness.Using the track stripping method to extract the skeleton line, and after the extraction of the road network frame to do post-processing.(2) Determination of vehicle driving direction and speed calculation of road sectionIn this paper, the mathematical plane geometry method is used to divide the taxi track points in different directions. Considering the huge amount of data, using method for quickly searching taxi track points to corresponding sections based on Grid. Using the arithmetic average method to calculate the average speed of the road section in a certain tme zone, and constructing the two-way speed signal sequence data.(3) Spectrum analysis of road speed signalIn order to effectively dig information from the frequency domain space, the fast Fourier transform of the two-way velocity signal sequence is carried out. The speed signal is transferred from time domain to frequency domain, analysis of potential cycle value by periodogram. Detection of urban road network speed variation cycle or rhythm variation characteristic of road speed signal. Based on the time and space distribution of the last week, we can see the following main features: Road network speed change cycle value of the larger sections are mainly concentrated in the train station, a specific traffic arteries, shopping district near the road; In the railway station, a particular traffic roads, shopping district and other sections of the reverse direction of travel is the size of the value of the larger differences; On the working day and on the rest day, the periodic value of the different travel direction is the opposite. Combined with periodic value information, the information of the speed change periodic value of the excavated road network includes the following contents: In the train station, shopping district,tourist attractions, specific traffic arteries near the road traffic flow is stable; Traffic flow is relatively stable in the direction of the regional which road traffic flow in different directions are different; the southbound traffic flow is more stable than the northward traffic flow.