Research On Urban Hot Spot Region Division And Path Guidance Based On Floating Vehicle GPS Data

With the continuous development of Intelligent Transportation System(ITS),people travel more conveniently and efficiently.Through the massive dynamic real-time GPS data acquired by floating car,data mining technology is used to study on spatial and temporal characteristics of residents’ travel,and analyze their travel patterns,spatial and temporal distribution.Spatio-temporal analysis of residents’travel has important research value for alleviating urban traffic congestion,carrying out more reasonable urban planning and formulating effective traffic control measures.In this paper,the floating car GPS data is taken as the research object.On the basis of data mining and processing,an algorithm for dividing hot spots in cities is designed.On this basis,the taxi route guidance algorithm is studied and validated.The research contents are as follows:1.Firstly,this paper preprocesses the acquired floating car GPS data,including removing noise data,redundant data,making up for historical missing data,extracting passenger’s boarding and alighting data,and using ST-Matching algorithm for map matching,which improves the accuracy of map matching,reduces the computational complexity,and provides data basis for later research,2.A K-Means clustering algorithm based on kernel density estimation is proposed to partition urban hot spots.Firstly,the density distribution of spatial data is calculated based on the Gauss kernel function,and the density threshold is set to estimate the density.The city is divided into high density area and low density area.Then the maximum density point of high density area is extracted as the initial centroid of K-Means clustering,and the distance between the extreme density point of low density area and the extreme value point of high density area is used to judge whether it is clustered separately.Iterative calculation is carried out to divide the hot spots in Qingdao.Finally,the algorithm is evaluated based on the contour coefficient of spatial similarity(AR])and clustering time efficiency.3.On the basis of the above research,the hotspot paths are extracted from the trajectory data in the hotspot area by similarity clustering based on LCS(longest space common subsequence),and the concept of hotspot path density function is proposed.Taking cruise destination congestion and taxi supply-demand ratio as preconditions,this paper designs a route guidance model with the shortest travel time of taxi to cruise destination as optimization objective,and validates the algorithm with no-load taxi passenger rate as evaluation index.