| Spatial accessibility has gradually become a research hotspot in the fields of geography,planning,and transportation,as an important evaluation index for residents’travel,transportation system and land use.With the development of mobile Internet and big data analysis,it has enabled researchers to study traffic accessibility from the perspective of spatial interaction by using massive amounts of data,like GPS trajectory data.On the other hand,private car ownership and residents’travel demand have been growing rapidly as the urbanization process continues.Traffic congestion has become one of the most critical issues which need to be handled seriously.Affected by traffic congestion and weather conditions,the travel time reliability of trips between the same OD will be fluctuated to some extent.Therefore,the accessibility study using a constant time threshold as the travel cost cannot demonstrate the impact of travel time reliability well.To date,this thesis has finished the following research:Firstly,we chose areas within the 5th ring road of Beijing as the study area and divided it into 500m hexagons to study the spatio-temporal variation of accessibility.The time interval and travel time of trips have been calculated and outliers in longitude and latitude have been preprocessed.Moreover,we applied the cumulative opportunity function as the accessibility measure and proposed algorithms to calculate accessibility results based on GPS trajectory data.The night-time light data was utilized to estimate the distribution of population density.From the perspective of healthcare facilities,we studied the spatio-temporal characteristics of healthcare accessibility and evaluate the matching degree between the accessibility and population density distribution.Besides,the autocorrelation method is used to analyze the spatial correlation of residents’healthcare accessibility to reveal the correlation between accessibility and the geographical location.Finally,we extracted and fitted the travel time of different OD pairs based on the GPS trajectory data.The finishing probability of OD pairs was calculated and the CDF(cumulative distribution function)is used to realize the conversion from the time threshold to the reliability constraint.Hence,the accessibility measure with reliability constraint was established.Based on this measure,the accessibility results under different reliability were calculated and the convenience degree of residents’life was assessed.Then,we studied the impacts of travel time reliability on accessibility distribution and distinguished the difference of accessibility results with/without reliability constraint.In conclusions,the spatial location has a significant impact on accessibility result.With the increase of the distance to downtown,accessibility dropped gradually but accessibility variation got bigger.With the improvement of reliability requirement,spatial accessibility reduced and it’s more sensitive to the reliability changes in the arterial road and of weekday travels.However,high reliability requirement will be beneficial in improving accessibility equity by reducing special variation of accessibility.In addition,from a theoretical perspective,ignoring the travel time reliability will underestimate accessibility level under low reliability constraints.When the reliability requirements gradually increase,the level of accessibility in different regions will be overestimated. |