Study On Traffic Detector Optimal Location For Dynamic OD Estimation

With the high-speed development of economic and the constant progress ofthe society, the city is expanding constantly and the amount of the automobileincreases sharply. Therefore traffic questions such as the traffic congest, trafficaccident etc perplexing the urbanite's daily life, and severe traffic condition havecaused social problem, financial losses and air pollution etc. ITS (IntelligentTransportation Systems) have already become the effective way to solve theseproblems. The countries all over the world invest a large number of financial andmanpower resources in carrying out the research in the field of ITS and havemade fruitful achievement.But, with broad application and deployment of ITS in our country, manyflaws have been uncovered. For example, absence of theory describing therealistic traffic behavior and lack of traffic data processing and mining method intraffic management department, which have badly restricted the development andoperation of ATIS(Advanced Traffic Information Systems) in our country. Aimingat the factors that restrict the development of ATIS, the research of DTA(Dynamic Traffic Assignment) and Dynamic Origin/Destination(OD) Estimationis valuable in theory and practice. And, in the application of DTA and DynamicOD Estimation, traffic detector mounted on the link is an important data source,the number and rationality of traffic detector layout project directly affect theperformance of model.This dissertation is supported by "Research of Fundamental TrafficInformation Collection Technologies and Optimal Project", which is the third partof the project of "Study on the Collection and Fusion Technology of theFundamental Traffic Information"launched by the nation. Based on the analysisof the factors that restrict the development of ATIS, we educe the concept of DTAand Dynamic OD Estimation, and bring forward the traffic detector optimalproblem under the DTA and Dynamic OD Estimation. Then, we describe the timedependent characteristics of traffic flow, and the relationship between OD matrixand link traffic volume in static/dynamic road network. Based on DTA baseddynamic OD estimation method, we propose the rules for optimal detectorlocations, and build the mathematical model. At last, we apply GA(GeneticAlgorithm) to solve the model. There are five chapters in this dissertation. Chapter 1: Introduction. This part first Introduce the research backgrounds ofthe dissertation and the development of domestic and international ITS. Then theauthor illuminates the research purpose and significance of this dissertation andprovides the main contents. Chapter 2: DTA and Dynamic OD Estimation. At first, we describe theconcept of traffic assignment and equilibrium. Then, based on the time dependentcharacteristics of traffic flow, we educe the research of DTA is necessary andimportant. At last, we describe the relationship between DTA and dynamic ODEstimation, and emphasize the Dynamic OD Estimation Method based on DTA. Chapter 3: Traffic Detector Optimal Location Model. First, we introduce theconcept of shortest route and valid route. Then, based on DTA based dynamic ODestimation method, we propose the rules for optimal detector locations, and buildthe mathematical model. Chapter 4: Solving the Traffic Detector Optimal Location Model. Based onthe form and restriction of Traffic Detector Optimal Location Model, weintroduce the GA (Genetic Algorithm) to solve the model. The, we apply thisalgorithm to compute a small-size network and a middle-size network. Chapter 5: Conclusions. Research contents of the full text are summarizedand the content of still needing to study and perfect is put forward. Traffic Detectors are sources of traffic data. In the process of DTA basedDynamic OD Estimation method, the number and rationality of traffic detectorlocated in the road network is very important for performance of the models. This...