Estimation And Analysis Of Urban Road Traffic Demand Based On Mobile Phone Positioning Data

Traffic demand analysis on prediction and assessment of traffic flow under different land use layout is the foundation of urban traffic planning. The Four-stages method is the traditional way to analyze traffic demand, but the travel survey data which this method needs usually takes much of man power and material resources, and the method belongs to the disaggregate model so itself also has some limitations. As the sensor technology, computer technology and communication technology development, a large number of new data is available, which provide us the opportunity to use the advanced technology to solve the traditional traffic problems. In this paper, we will put forward a new method that use mobile phone positioning data as the basic data to analyze road traffic demand.Based on the traditional traffic demand forecast method of four-stage model, this paper puts forward a method based on mobile phone positioning data among urban residents and integrates multi-source traffic data to assign the traffic flow to the urban road system, and this paper takes Shenzhen as an instance of the method, estimates and analyzes the road traffic demand. According to the analysis of the citizen’s stay and move pattern, this method firstly uses short-term regular sampling of mobile phone positioning data instead of the traditional questionnaire survey data to extract the urban residents travel path, and then takes cell station as the travel unit, extracts the travel Origin-Destination(OD) matrix that residents travel between cell stations also known the traffic distribution between the cell stations. Considering the urban residents travel mode, the inhabitants transportation way is divided into personal transportation and public transportation, which use a month of floating car data and public transportation smart-card data to acquire the historical trajectory set on them severally, and when extracting the travel path on the public transport system the public transport transfer is considered. According to the historical trajectory set, any empirical probability travel path between OD pair of site can be known, no matter in urban road network or bus and subway station in public transport system. Combining historical trajectory set information, traffic modal split model and traffic flow assignment model can be established, the traffic distribution between the cell station can be split on traffic modal and assign into general road network and public transportation road network by their model, finally traffic flow on each road is calculated in every time period, so as to realize the estimation of the whole urban road traffic demand. Finally we use the bayonet monitoring data and smart-card data to validate and analyze the estimation results of the traffic demand.This method offers a good joint to integrate the mobile phone data and other urban traffic data, which can take advantage of the high coverage property of mobile phone data and accurate location information, high sampling frequency properties of urban traffic data such as GPS. This method also gives a novel insight and thought to use the multiple-source traffic data in the practical application.