Research On The Impact Of Bace Station Layout On The Extraction Accuracy Of OD Based On Cellular Signaling Data

With the improvement of residents’ living standards,people’s travel frequency is gradually increasing,and the way of travel is increasingly diversified.In the face of the huge demand for transportation,transportation planning and demand management are facing great challenges.For a long time,traffic planning has been based on the survey data of residents’ trips for prediction.This method has high survey cost,high demand of participants,and inaccurate subjective recall of respondents,so the quality of data and the efficiency of the survey cannot meet the increasing planning requirements.With the rapid iteration of information technology and the in-depth development of smart city,it is possible to monitor residents’ activities based on mobile signaling data,which can effectively improve the efficiency of demand prediction.Travel endpoints and OD information are the basis of followup monitoring and analysis.Therefore,how to improve the accuracy and location accuracy of endpoints and OD is of great importance for future applications in the field of transportation.This study focuses on the Angle layout of base stations,analyzes the impact of Angle layout on the effect of OD extraction,and takes it as an influential factor in the study of OD extraction,so as to achieve efficient extraction of travel endpoints and OD.This paper firstly designed and carried out an empirical experiment of volunteers in G City,collected volunteers’ log data,GPS track data and mobile phone signaling data synchronously,statistically analyzed the data amount and time interval of mobile phone signaling data,and constructed an integrated simulation platform of "communication-traffic" according to the measured mobile phone signaling frequency.Secondly,aiming at the shortcomings of existing endpoints and OD extraction,this paper proposes a two-stage optimization method for travel endpoints and OD identification.The first stage,with the improved recognition algorithm,the depth of the forest endpoint fusion cell phone signaling data and point of interest(POI)data,four categories,a total of 10 selected attributes,combined with the travel log data and GPS trajectory data empirical comparison evaluation,the final results show that the depth of the forest algorithm to identify the endpoint accuracy and multiple recognition effect is best,89.3% and 5.6%,respectively.In the second stage,an endpoint position optimization algorithm considering the base station Angle layout is proposed,and the OD extraction accuracy is improved by using the base station direction Angle information.The results show that the average distance error is reduced by 38 m by using the direction Angle information,which is 14.5% higher than the original algorithm,and the OD identification effect is 4.4% higher.Finally using "communication and transportation" simulation platform,the travel OD optimization algorithm under different base station density extraction effect,the results showed that the distance error optimization effect is the best density range per square km area of base station number is less than 50,the optimization effect is 84.2 meters,OD precision optimization effect is the best density range is 150-200,Accuracy improved by 4.99%.The two-stage OD identification method effectively improves the extraction efficiency of travel information,is expected to save the cost of investigation on residents’ trips,and provides strong data support for traffic planning,management and monitoring.