Study On Estimation And Prediction Methods Of Travel Time Distribution In Urban Road Networks

This paper mainly studies the estimation and prediction method of travel time distribution,and aims to provide real-time,accurate and reliable traffic information,and make reasonable travel plans for travelers to reduce travel time and travel delay so as to alleviate traffic congestion and reduce environmental pollution and energy consumption.Travel times in urban road networks are highly stochastic.However,most existing travel time estimation methods only estimate the mean travel times,while ignoring travel time variances.To this end,this paper proposes a robust travel time distribution estimation method to estimate both the mean and variance of travel times by using emerging low-frequency floating car data.Different from the existing studies,the path travel time distribution in this study is formulated as the sum of the deterministic link travel times and stochastic turning delays at intersections.In the study,the travel time is simulated as a deterministic variable,and the concept of degree of central tendency is proposed to estimate link travel time.The delays of the different turning movements(through,right-turn,and left-turn)are simulated as random variables and subject to a Log-normal distribution,which is consistent with the observed travel time verified by the chi-square test.In addition,considering the low sampling rate of the floating car data,a weighted moving average algorithm is further developed for a robust estimation of the path travel time distribution.A real-world case study in Wuhan,China is carried out to validate the applicability of the proposed method.The results of the case study show that the proposed method can obtain a reliable and accurate estimation of path travel time distribution in congested urban road networks.For the heterogeneous data source with different data quality and network coverage generated by interval and point detectors,a heterogeneous data fusion method is proposed to estimate travel time distributions by fusing these data.In the proposed method,link travel time distributions are first estimated from point detector observations.The travel time distributions of links without point detectors are imputed based on their spatial correlations with links that have point detectors.The estimated link travel time distributions are then fused with path travel time distributions obtained from the interval detectors using D-S evidence theory.Based on fused path travel time distribution,an optimization technique is further introduced to update link travel time distributions and their spatial correlations.A case study was performed using real world data from Hong Kong and showed that the proposed method obtained accurate and robust estimations of link and path travel time distributions in congested road networks.For the problem that the classical KNN model can not predict travel time distribution,a stochastic KNN model is proposed to predict the travel time distribution of multiple periods in urban roads.The model overcomes the defects of some current models,such as only using real-time data,predicting single-period travel times,and can not predicting the distribution.In the proposed model,four similarity measures are used to find the nearest neighbor of the distribution.It not only predicts the travel time mean and variance,but also predicts the type of travel time distribution,which can give an accurate travel time interval.Based on the automatic vehicle identification data from Hong Kong,the proposed stochastic KNN model was cross-validated,and the mean absolute percentage error of the distribution,the coverage of predicted interval and the cumulative distribution error of the distribution prediction were analyzed,which proved the effectiveness and robustness of the proposed method.The sensitivity analysis was carried out from four aspects:the number of nearest neighbor,the number of adopted time intervals used in the prediction,the similarity measure and the type of predicted distribution,which proves the reasonable of the parameters used in the experiment.For the problem of different characteristics and applicability of different prediction methods,a combined prediction method based on D-S evidence theory is proposed.Three common models are selected as the sub-prediction model,and the travel time distribution of each sub-model is recombined to get a better travel time distribution.Different from the traditional linear combination model,the D-S evidence theory is used to fuse the travel time distributions of multiple prediction models,so as to achieve the purpose of combined prediction.At the same time,the combined model also reduces the negative influence of low precision on the prediction results,and can make use of various model information.The results showed that the combined prediction method based on D-S evidence theory is more scientific and effective than the single prediction method.