Research On Impact Of The Location Frequency Of Cellular Data On The Extraction Accuracy Of Transport Modes

Travel information has always been an important basis and supporting data for traffic planning and traffic policy.It is an important part of transport modes for travel information obtaining.Mastering the current travel structure and its development of transport modes is the basis for decision-making in traffic planning and traffic management.With the development of 4G mobile communication technology,the increase in the location frequency of mobile phone signaling data has brought opportunities for the fine-grained extraction of travel information.How to extract individual fine-grained travel information such as transport modes from mobile phone signaling data and evaluate the accuracy of extraction is an innovation in traffic investigation methods.A travel experiment is designed in this paper to collects real mobile phone signaling data.A traffic-communication integrated simulation platform is constructed to generate simulated mobile phone signaling data,combining real traffic environment and 4G mobile communication network principles based on individual real travel data.With the advancement of mobile communication technology,the location frequency of mobile phone signaling data has been improved.The basic characteristics of real mobile phone signaling data are analyzed,and the probability theory knowledge is used to explore the distribution law of location frequency,which is loaded into the simulation platform to simulate mobile phone signaling data of different location frequencies.Based on the mobile phone signaling data,a random forest-based transport mode identification model is constructed to use the collected mobile phone signaling data as the basic data source to identify the traffic mode and verify the effectiveness of the algorithm.In order to evaluate the influence of different location frequencies on the accuracy of transport mode recognition,the relationship between the location frequency and the accuracy of identifing transport modes is quantified based on mobile phone signaling data.It is found that when the location frequency is higher,that is,the time interval between mobile phone signaling data generation is smaller,the recognition accuracy of transport modes is higher;however each mode of transportation has different sensitivity to the location frequency,among which the bus and the car are more sensitive.Through analysis,with the increase in the location frequency,the recognition accuracy of the bus and the car is increased by 36% and54%,respectively increasing to 70%,89%.When the location frequency is less than 40 s,the recognition accuracy rate is still improving,but the increase is small.Compared to bus and car,the recognition accuracy of walking and electric vehicles has a small increase with the increase in location frequency.