Research On Seeking Behavior Analysis And Path Planning For Taxi Based On GPS Data

In the process of taxi operation,traveler’s requirement information has an important impact on the efficiency and profit of taxi.At the same time,the time-varying nature of traffic environment makes it difficult for drivers to seek passengers effectively.In order to reduce the occurrence of such situations,it is particularly important to recommend real-time and effective passenger-seeking solutions for taxi drivers.Therefore,this paper establishes a two-stage recommendation model for taxi passenger-seeking scheme.The model dynamically recommends the scheme to improve the efficiency of driver seeking passengers.The main contents of this paper are as follows:(1)GPS data matching and analysis of traveler’s preference informationIn order to grasp the travel temporal and spatial regular pattern of travelers,this paper conducts data cleaning and data analysis based on the Beijing Taxi GPS historical data,and the effective information is extracted from the data.And the extraction algorithm of the origin destination(OD)and the matching algorithm of the taxi OD data are established.Based on the analysis of OD data,the taxi demand and operation time information are obtained,and the characteristics of urban residents’ travel preferences is analyzed,which provide data support for determining the taxi-seeking target in the following paper.(2)Region division and calculating method of searching indicators based on VoronoiIn order to accurately recommend the search target for taxi drivers,the O points are clustered based on the taxi OD data.According to the Voronoi diagram method,the research area is divided to the driver search target area set based on the cluster center coordinates.According to the area statistical data,the calculation model of area passenger-seeking indicators is established.For the selection of area passenger-seeking indicators,this paper establishes the driver’s passenger-seeking characteristic model based on kernel function,which reveals the change law of the attraction of the target area and the distance from the driver to the target area,and the model studies the influence of the characteristics of the passenger-seeking area on the driver’s passenger-seeking strategy.This part does the basic work for the recommendation of the passenger-seeking scheme in the following paper.(3)Dynamic search method based on uncertain travel timeThis paper establishes a two-stage passenger-seeking scheme recommendation model for taxis,which are the target area recommendation model and the target point recommendation model respectively.By extracting the data of various searching indicators from GPS data,the paper calculates the seeking recommendation model,and taking the seeking scheme with the highest recommended value(that is,the optimal seeking scheme)as the initial scheme of the passenger-seeking.Based on the recommendation model,this paper proposes a dynamic seeking method based on uncertain travel time.By predicting the traffic speed of the road sections in each path,the road traffic indicators and the empty travel time are estimated and calculated.And then the recommended value of each target point is calculated and updated.At last,the driver’s optimal seeking target point is adjusted in real time.Based on the travel characteristics obtained from the analysis of the taxi GPS data,this paper establishes a passenger-seeking recommendation model for the taxi drivers to clarify the passenger-seeking target.Finally,the travel time of taxi passenger-seeking section is estimated,which is used for real-time updating and calculating the searching target of the driver.The goal of passenger-seeking recommendation model is to improve the efficiency of taxi searching for passengers.