| The travel route data of motor vehicles contains a large amount of traffic spatiotemporal information,by extracting and segmenting the travel routes of motor vehicles,the travel situation of motor vehicles can be systematically reproduced,thereby providing data support for the analysis of traffic structure and traffic planning and management.Most of the previous studies were based on public transportation big data to deduce the operation status of all motor vehicles in the city,but ignored the differences between public transportation and private cars.In addition,most of the research was based on the more mature intelligent transportation system in large cities,which makes it difficult to directly transfer some methods and theories to small and medium-sized cities.Therefore,using the traffic bayonet data generated by the smart transportation system in small and medium cities,this paper proposes a method of vehicle route reconstruction,which reproduces the real travel conditions of motor vehicles to the greatest extent and combines the idea of clustering to identify commuter vehicles and analyze travel characteristics to provide a basis for traffic management and policy formulation.The main contents of the thesis are as follows:(1)Traffic bayonet data preprocessing.Firstly,a preprocessing method of traffic bayonet data was proposed.Then,the point data of traffic bayonet equipment were modified,including the actual point and virtual point.The bayonet network topology was constructed in the spatiotemporal data of road network,and the filtering method of road travel time was proposed combining with the existing road travel time processing method.Finally,the distribution and characteristics of traffic bayonet data were analyzed.Finally,the longest common subsequence method(LCSS)is used to judge the spatial similarity between the reconstructed path and the real path,which verifies the accuracy of the algorithm.(2)Vehicle path reconstruction.Analysis of the disadvantages of the existing vehicle path reconstruction method,puts forward the path reconstruction method based on the monitoring data,the spatial-temporal data network as well as the motor vehicle bayonet car data,completed the one-day travel path of motor vehicles and the use of dynamic time threshold,completed the carving path to exist at the same time the missing part.An alternative route decision algorithm based on KSP algorithm and TOPSIS algorithm was proposed to reconstruct the single-day travel routes of motor vehicles.(3)Commuter vehicle identification.Firstly,the peak travel time was determined according to the vehicle travel situation in the city.and a set of methods for determining vehicle travel OD points and calculating OD stability were proposed.Secondly.the travel characteristics of motor vehicles were extracted.the dimension of feature index was reduced by factor analysis method and the common factors are determined.Then.the k-means algorithm and FCM algorithm were used to identify the commuter vehicles.and the results were compared with statistical methods.Finally,the travel characteristics of the commuter vehicles were analyzed.At the same time,the problems existing in yueyang city center are analyzed and the corresponding improvement suggestions are put forward. |
