Research Of Mobile Traffic Detection Method Based On Smartphone

Mobile traffic detection technology, which has advantages of broad coverage,strong real-time, low operating costs, can effectively compensate for the defects of thestationary traffic detection technology. Although mobile traffic detection technologybased on floating cars have made satisfactory progress in data acquisition, processing,application, the measurement accuracy is limited by the proportion of the number offloating cars of that of vehicles in the road network. In order to fundamentally solvethis problem, this paper researches a kind of mobile traffic detection method based onsmart phones and try to design the proposed new applies to the functional modulesthrough in-depth study in data preprocessing, map matching, traffic modeling fittingand to improve the proposed new applies to the functional modules of the method.With the development and progress of the smart phone positioning method, thispaper expects that the main method of the smart phone positioning will be GPSpositioning. In the data preprocessing stage, this paper researched the methods oferror correction and coordinates transformation in GPS and gave a specific calculationprocess.After reading and analyzing extensive literature on map matching algorithm ofvehicle navigation system, this paper points out the differences of the functionalrequirements between map matching algorithm in mobile traffic detection and that invehicle navigation system. To complete the map matching of the smart phone mobiletraffic detection method based on smart phone, this paper designs a kind of fuzzy logicalgorithm making use of road topology information based on the analysis of smartphone users’ four kinds of traffic behavior.Flock mode, which is one of relative motion of moving point objects, caneffectively simulate smart phone users riding in the same car. First, this paperintroduces the concept of Moving Point Objects (MPOs) and Relative Motion (REMO),and focuses on the analysis of Flock pattern, which is one of the REMO patterns.Considering the special distribution laws of mobile phone users in road traffic, thispaper proposes an identification model which can identify the number of mobilephone users in the same vehicle, based on r-first algorithm in Flock pattern.To verify the effectiveness of the core function module-the traffic parameters estimation, I programmed part of the core functions, and use the VISSIM, which is akind of traffic simulation software, to design the simulation experiment. Theexperimental results show that in simulated conditions, the relative error of thesegment average speed calculated by our method is only4.35%, which achieves thedesired accuracy requirements.In addition, part of the results, for example, Flock model-based smart phone useridentification algorithm, can inspire other researchers who focus on mobile GIS spatialanalysis studies or on mobile GIS public application.