Study On GPS Taxi Empty-Load Speculation And Analysis Of Driving Behavior

Taxi has long been one of the important factors that affect the planning and management of urban traffic.The research on taxi GPS at home and abroad is summarized,including OD travel,taxi operation,traffic behavior,traffic state estimation and so on.In the above study,in addition to traffic estimation,the rest of the research can not be separated from the vehicle information.However,only a small portion of the collected taxi GPS data contains vehicle information in the collected taxi GPS data,which is not closely connected or the signal is poor.By consulting a large number of data,there are few studies on this aspect,but this part is a problem that needs to be solved urgently.In addition,the mortality rate of traffic accidents caused by illegal driving has been high in recent years,which has caused serious damage to people’s lives and property and has become an urgent problem for the government.Based on GPS data,this paper mainly carries out two aspects of research.First,it is estimated that the taxi is empty-load state through GPS.The completion of missing data,provide data support for the follow-up study.Second,the statistical analysis of illegal driving behavior and the development of visual system.Through analysis,we can perceive the temporal and spatial distribution rules of two kinds of illegal driving behaviors,such as operation vehicles,speeding and retrograde traffic between different roads and different functional areas,and provide intuitive and reliable data support for urban traffic management decision-makers.Firstly,the original data are cleaned,coordinate conversion and map matching.Secondly,a two stage model is proposed to speculate on the no-load information of taxis.By taxi operation characteristic that,between passengers and two vehicles parking sections for passenger track.Therefore,this paper will speculate that the no load and loading capacity of taxis should be focused on the first stage,that is,finding the location of passengers’ getting up and getting off,specifically the integration of ResNeXt and BLSTM to form a new network structure for speculation.The second stage needs to be combined with map information for global correction.Finally,the complete taxi passenger status information is formed.Thirdly,this paper puts forward a method of dividing the level of violation.Through this method,the speeding and retrograde behavior is extracted from Xinjiang GPS data,and the distribution and change rule of vehicle and illegal vehicle near the working days and rest days near different functional areas are analyzed.Finally,the driving characteristics of a very dangerous driver,grade 4,were discussed.Finally,the taxi violation driving behavior analysis system is designed.By using JAVA and Gao de map API programming,the visualization of taxi operation vehicle analysis,speeding vehicle analysis,retrograde vehicle analysis and single track analysis is realized.