Research Of Bus Travel Time Prediction Based On Similar Road Segmentation

The positive development of public transport is a more effective way to alleviate the pressure of urban traffic.In order to increase the concern of public transport passengers and improve the service level of urban public transport,many inland and abroad scholars have made a large number of studies on prediction of bus arrival time,the previous literature pay attention to the bus forecasting algorithm for a single bus line information,due to bus departure time interval,a single line of information not enough to reflect the real-time traffic conditions,resulting in real-time prediction results are not high.The previous scholars have little research for the bus station delay of bus arrival time,the bus entry the station time and exit the station time is not clear,affecting the accuracy of the forecast results.In view of the above shortcomings,this paper proposes a multi-station similar road segment forecasting method,based on the classification of similar sub-sections,taking into account the characteristics of traffic conditions similar to the use of multi-line bus information Kalman filter algorithm to predict bus station Travel time,and the bus station docking process done in-depth analysis,the use of clustering algorithm to access the site time to cluster,get the site of the corresponding time of the docking time,and ultimately predict the bus on a road running time.The specific work of this paper is as follows:(1)This paper analyzes the influencing factors of the transit time between the bus stations,defines the concept of similar road sections,and divides the bus lines.The similar sections have similar road traffic capacity.In the same period,the speed of the bus traveling on different routes is similar to each other,the Kalman filter algorithm use different routes information to predict the traveling time between the bus stations and improve the real-time performance of the bus arrival time prediction results.(2)A similar section of the road contains a number of bus stations.In order to research the bus station delay time,need to analyze the bus inbound and outbound process in detail,this paper defined the bus inbound and outbound time point,the bus began to slow down into station to speed up out of the station is bus service cycle,analysis of bus docking-regularity on workdays and weekends,use K-means algorithm for its clustering,get the docking time of the site corresponding time period.(3)In this paper,the experimental line of Shenyang 162 line and 299 line is selected as the prediction algorithm model of the paper.Two lines each draw a similar section of the basic data collection and experimental.The experimental results verify that the algorithm model has better real-time and reliability.