Bus Route Optimization Based On Taxi GPS Trajectory Data

As the economy grows,people are increasingly pursuing a comfortable life.In terms of travel,people pay more attention to enjoyment and convenience,which makes private cars popular and become one of the most important ways for residents to travel.However,with the gradual increase in the number of private cars owned by residents,urban problems are becoming more and more serious.This requires us to conduct a detailed analysis of urban traffic problems and propose solutions to alleviate this situation.Transportation methods such as buses and taxis are important links in current urban transportation.However,due to problems such as transfer and delay,the attraction of the bus to residents is reduced,making urban residents more inclined to taxi.How to make the public transport system more suitable for residents’ travel needs has become a problem that has received much attention.The trajectory data of the taxi contains a lot of information.By analyzing this information,it can sense the road conditions of the city roads,and dig out the hot areas in the city and the travel rules of the residents.In combination with the existing bus network in the city,the defects of the city bus system can be unearthed,and the bus line network can be improved to better meet the passenger demand.This paper combines the data of the taxi city of Porto to analyze and improve the existing bus routes in the city,so that the bus routes are more convenient for passengers to travel.The main contents and results are:1.Research on clustering algorithm using groupingWhen the amount of data is large,the traditional DBSCAN algorithm and some existing improved algorithms are not too efficient.Aiming at this feature,this paper combines the characteristics of taxi GPS data and improves the GDBSCAN* algorithm based on the packet-based DBSCAN algorithm.The algorithm converts some distances that need to be repeatedly calculated in the clustering process into distances that can be used multiple times,which improves the clustering speed.2.Research on intersection identification methodFor the sparsely sampled GPS trajectory data,some current intersection recognition algorithms identify the characteristics of low accuracy.This paper proposes a method for extracting the frequent occurrence of parking events in urban roads by using vehicle parking points,and using the subsequent points to detect intersections.This improves the accuracy of intersection detection using such data.3.Planning of bus routesIn this paper,the taxi GPS trajectory data is used to study the traffic division,road condition analysis,passenger travel hot zone analysis and bus route planning methods,and combined with the taxi data in Porto area,the city’s existing bus network has been carried out.Analysis,improvement and planning of bus routes for routes where a large number of taxi trips are gathered.Finally,the improved route was evaluated using a number of indicators,and theresults showed that the improvement achieved the convenience of passengers.