Identification And Analysis Of Connectivity Reliability And Critical Links Based On Actual Road Network

The state of traffic is an important indicator of the development level in city.As an important service carrier to meet travel demands of urban residents,a stable,efficient and reliable road traffic system is not only the wish of traffic travelers,but also the ultimate goal of management and builders.Traffic reliability is a critical indicator for measuring the quality of road network services and assessing the status of transportation systems.In order to better quantify the reliability of the urban road network,we choose indicator of connectivity reliability to study the actual road network in Chengdu based on trajectory dataset from Didi.The specific research contents and conclusions can be summarized into the following four aspects.1.In order to establish topology structure based on the actual road network,we use ArcGIS to process OSM?OpenStreetMap?road network data.According to the actual connection situation,the road network is converted into a directed graph represented by a sparse matrix.2.We propose a method for converting the trajectory data without velocity data into the relative velocity of the links.After finishing transforming coordinate system and map-matching on the trajectory data,the average velocity between the trajectory points is obtained by calculating the straight distance.Then we use a"voting"method,considering the velocity limit condition,to calculate average speed of the road segment.We consider the road velocity limit for each road as 95 percentiles of its velocity.Then the ratio between actual velocity and road velocity limit is defined as relative velocity.3.In the calculation of connectivity reliability,the concept of strong connectivity is introduced,and we propose a dynamic screening indicator k.A link will be connected only when its relative speed is higher or equal to k during a certain period.In order to quantitatively analyze the connectivity of the road network,the Tarjan algorithm based on DFS is used to identify strongly connected components in the road network.It is found that the size of the largest strongly connected components will decrease rapidly in the vicinity of a critical state,which is defined as the point for calculating the connectivity reliability index6)₍?8? Finally,through the calculation of actual data,the difference of connectivity reliability in working days and weekends in Chengdu local road network and its relationship with the average speed of road network are explored.4.Based on the model of connectivity reliability,the definition and discriminant conditions of critical links are proposed,and the trend of connectivity reliability under single-segment and multi-segment conditions is explored.By analyzing the frequency and time of occurrence of critical links,it is found that the temporal distribution of critical links presents a certain regularity.It is found that different links of roads affect road network connectivity in different time and critical links appear regularly during the same period on different dates.We propose a complete trajectory data processing method,a new traffic reliability model to quantify the connectivity reliability of the corresponding region and a method to identify critical links.Our research provides a new viewpoint of analyze connectivity reliability,which helps the traffic management department to identify the actual road network status and propose low-cost traffic management schemes.