Research On Freeway Travel Time Reliability Evaluation Method Based On Mobile Phone Data

With the sustained and rapid development of China's social economy and freeways,motor vehicle ownership continues to maintain rapid growth.At the same time,people's travel demand is increasing day by day,and the traffic volume of freeways is also increasing rapidly,resulting in the deterioration of freeway operation.Therefore,improving the operation efficiency of freeways through the analysis and evaluation of traffic condition and level of service is a project worthy of research.However,currently scholars at home and abroad often utilize non-probabilistic index such as saturation,average speed and delay in the evaluation of freeway operation.Although these index are easily available,they cannot reflect the randomness and volatility of the freeway operation in terms of temporal and spatial dimensions.Therefore,it is necessary to establish the probabilistic evaluation index to analyze the randomness of freeway operation,and one of the most effective methods is the travel time reliability.In recent years,mobile communication technology is rapidly developing,and mobile phone has become an indispensable communication tool in people's daily life.Traffic information collection technology based on mobile phone is also gradually emerging.The information of mobile phone data is rich in content and huge in amount.Also mobile phone data is low-cost,high efficiency and high accuracy.It has incomparable advantages and broad application prospects.Therefore,freeway travel time reliability evaluation method based on mobile phone data is researched in this paper.Firstly,in order to fully understand the concept of travel time reliability and the generation mechanism of mobile phone data,the related concept of travel time reliability is reviewed,and the mobile phone handoff location technology and traffic information extraction technology is comprehensively analyzed.Combining mobile phone data with travel time reliability can extend the application of the data.Secondly,the comprehensive travel time reliability evaluation system for freeways is established.The evaluation system is structured and has overall index.It includes three parts namely the target layer,the criteria layer and the index layer.The index layer is supplemented by probabilistic index,statistical index,and TTI(Texas Transportation Institute)index.Each kind of them is quantified to provide a comprehensive,scientific and reasonable evaluation of the freeway travel time reliability.The nextly,the probability distribution model for travel time of freeway links is established.The empirical analysis method is used to analyze the travel time volatility from temporal and spatial dimensions.Utilize mobile phone data to conduct curve fitting and parameter estimation of six common travel time probability distribution models such as lognormal,generalized extreme value,inverse Gaussian,Gamma and etc and a travel time probability distribution model based on lognormal distribution is established.Then,the freeway travel time reliability evaluation model is established.The travel time threshold and the corresponding level of service are divided based on statistical analysis of travel time and the relationship between speed and volume.Establish the travel time reliability evaluation model for link based on lognormal distribution(LOGN-L),and considering the correlation between links,the travel time reliability evaluation model for path(LOGN-P)is built,and the accuracy of the evaluation model is verified based on mobile phone data.Finally,a case study based on the evaluation model proposed in this paper is conducted.Utilize probabilistic index,statistical index and TTI index to evaluate travel time reliability of freeway links based on mobile phone data collected from Shanghai-Nanjing freeway,and bottleneck links of freeway operation can be found.Utilize probabilistic index to evaluate the travel time reliability of freeway path,and it is proved that the correlation between links is an important factor that must be taken into accout.