Research On Network Connectivity Reliability Based On The GPS Data And GIS Technology

Presently,the network connectivity reliability is an important factor to reflect the performance of network structure.It’s easy to access the speed by GPS floating car data widely used in travel time estimation.This paper tries to expand the scope of application of GPS floating car data and combine it with the research about road network connectivity reliability.First,the paper analyzes the characteristics of floating car data and the principle of acquisition system,summed up the four shortcomings:low precision,lack of information,large sampling interval and abnormal data.Taking the floating car data in Chengdu as an example,this paper illustrates the information contained in the original data of the floating car,and selects the partial road network area of Chengdu as the research object.According to the characteristics of the data and the specific situation of the road network,this paper summarizes the 7 principles in data selection,which lis helpful for the next research.Then,the filtered GPS data is matched to the electronic map and divided into three parts for matching with reconstructing matching path:the first part is determining the error region and the candidate road set.According and the root mean square error and the circular error probability of the GPS data,the candidate region radius is 41.7m;second,divide the series of GPS matching points into three part:historical point,matching points and current point.Consider the relationship between the distance and angle of the GPS point with the candidate links called road geometric factors,and use the connected factors of adjacent floating car data points,to constructs a model matching strategy based on delay map;third is model complement part.From the relationship between the speed of GPS point with road speed limit and the identification of elevation and ground road,to perfect the matching model.Finally,on Visual Studio 2012 programming environment,use C#language and ArcEngine embedded GIS component library to visualize the map matching,and realize a series of operation from the import of GPS data to complete matching and export the results.Then,based on the map matching results and network topology information,this paper summarizes two kinds of method of calculating travel time:one is the estimation method of travel time based link distance which is considered the intersection delay;the orther one bases on harmonic mean velocity and consider the velocity characteristics of different sections of link.In order to make full use of the data to get effective and reasonable travel time,according to the idea of different parameters in the same population,this paper combine the two methods aboved.Finally,these two paths are randomly selected,compared the calculating result with the actual travel time,which was found that the calculating time of the road is consistent with the actual traveling environment.Finally,based on the relationship between speed with time and density,the paper perfect the concept of link connectivity by introducing Underwood exponential model and Greenberg logarithmic model to distinguish the change of traffic density.Next,on the one hand,use path searching algorithm to seek the feasible path between the OD pair,on the other hand,calculate the connectivity reliability of path with the theory of series-parallel system.Then,take travel time and length of road as weight coefficient to calculate the connectivity reliability of OD pair on network.Finally,take the importance of both ends and connected links on each OD pair as weight coefficient to analysis the connectivity reliability of the entire network.