Travel Characteristics Recognition Based On Data Fusion

Traveler's travel purpose information is one of the key criteria for urban traffic planning or optimization.The existing information collection of bus travel purpose relies mainly on manual acquisition.This method requires a lot of manpower and material resources,and often uses manual survey methods.A small sample survey is conducted,so the obtained results depend on the reasonability of samples.As a middleman between urban public transportation and private transportation,the travel purpose of taxi rider can effectively assess the connection between urban public transportation and private transportation.With the development of information technology,more and more operational data of vehicles can be automatically obtained.Therefore,how to combine the acquired information data and passenger travel information obtained by certain technical means has become a research hotspot in recent years.And gradually good results are achieved in practical applications.Firstly,this paper constructs a travel feature extraction model for taxi passengers.Combined with the operation data of the taxi,the characteristics of the taxi passenger travel distance and travel time are obtained by recording the change of the on-board field value in the passenger operation data.Secondly,the taxi passenger travel destination recognition algorithm and model based on the combination operation of taxi operation data and online POI data fusion are proposed.Summarizing the existing research methods,it is pointed out that only using travel characteristics will lead to low recognition accuracy,especially the problem of different travel destinations with similar characteristics,By integrating taxi operation data and online map POI data,the acquisition of taxi passengers is achieved.This method can effectively avoid the problem that different travel destinations cannot be accurately identified by combining machine learning and POI categories near the drop-off point.Compared with the travel feature alone,the proposed method can effectively improve the recognition accuracy.Finally,the accuracy of the model identification was verified by actual survey data of Chengdu,and the travel purpose identification was realized for the actual operation data of Chengdu taxi.The verification results show that the accuracy of model recognition after the integration of POI data is improved by 12%,and the proposed method can be applied to the identification of the actual destination of large-scale taxi operation data.The traveler information acquisition model constructed in this paper can effectively analyze the travel destination information and provide auxiliary decision information for urban traffic planning and optimization.