OD Demand Estimation Based On Trajectory Turning Behavior

Origin-destination(OD)demand estimation is a classic problem of interdisciplinary research of applied mathematics and transportation.Considering the influence of trajectory turning behavior on data analysis and model formulation,this thesis establishes an OD demand estimation model based on trajectory turning behavior.The trajectory turning behavior was considered in the sensor location problem,and the model was established to make full use of the observed data to estimate the OD demand,and the spare sensor location model was established based on the failure of the sensor.The following is a summary of each chapter:In chapter 1,the research background and the significance of OD demand estimation are introduced,the research status of the sensor location,OD demand estimation,and turning behavior are summarized.Moreover,several relevant preliminary knowledge is presented.In chapter 2,the generalized least squares model of OD demand estimation is improved based on trajectory turning behavior.It is proved that the consideration of trajectory turning behavior is helpful for reducing the estimation error for OD demand estimation.The effectiveness of the proposed model is validated on a small transportation network.In chapter 3,an initial sensor location model is established and an initial sensor location scheme is obtained by using the branch-and-bound method to solve the model.Then,a sensor location optimization model considering trajectory turning behavior is developed.The sensor location scheme with the least number of links equipped with sensors is obtained by solving the sensor location optimization model based on trajectory turning behavior through genetic algorithm under the condition that the unique solution of OD demand could be obtained.Finally,numerical examples are carried out to validate the effectiveness of the proposed models on two transportation networks.In chapter 4,under the condition that the transportation network has been equipped with sensors in chapter 3,the case that the possible failure of sensors is considered.According to the OD coverage principle,the maximum flow truncation principle,and the minimum path principle,the standby sensor location models are developed,and the obtained observable link flows are used to estimate the OD demand.The effectiveness of these models is verified on the two transportation networks,and the performances of the models are compared and analyzed.Experimental results show that the sensor location model based on OD coverage principle performs the best.The fifth chapter summarizes the research content and results of this thesis,and puts forward research prospects in further studies.The paper has 14 figures,19 tables,and 99 references.