OD Matrix Estimation Based On Turning Flow & Dynamic Traffic Assignment Based On Micro-Objects

The obtaining of Origin-Destination matrix and traffic assignment are two core in traffic demand forecast. The estimation of OD matrix from traffic flows is a significative work. The constraint of existing OD (Origin-Destination) matrix estimation methods is usually based on road segment traffic flow. It can leads to a solution space of high-dimensions, since the number of roads is much fewer than the OD pairs. Because of this, the error of the estimated OD matrix is large. Traffic assignment is the destination of traffic demand forecast, and plays important part in OD estimation. Dynamic traffic assignment is better than static traffic assignment because it considers the time variety of traffic networks. The existing dynamic assignment methods are on the basis of macro-flows, the mathematical programming problems of these methods are hard to solve and hard to explain.The main contents and contributions of this thesis can be summed up as follows.1. Some researches are done on several important problems in the traffic assignment method. This paper gives the traffic networks and OD points' description which are suitable for Urban Mixed Traffic. A shortest path algorithm based on User Decision Factor is given. User Decision Factor means that the shorter a road sect is away from the driver, the more important the road sect's cost when the driver chooses his path.2. Turning traffic flow reveals more traffic information than road segment flow. Using turning flow in the estimation of OD matrix can effectively reduce the dimensions of the solution space, and highly improve the estimation result of OD matrix. This paper gives an OD matrix estimation method based on turning flow, including the traffic assignment method based on turning flow, OD matrix estimation method based on turning P-matrix, and gives a compare between this method and the OD matrix estimation method based on road segment flow though theoretical proving and an example. This method can highly reduce the dimensions of the solution space, and improve the estimation result of OD matrix.3. This paper also gives a Dynamic Traffic Assignment Method based on micro-objects. This method is on the basis of Micro-simulation system. It avoids the disadvantages of static traffic assignment, the discontinuity of traditional dynamic traffic assignment based on macro-flows and the mathematical programming used in macroscopical method.4. As the chief engineer, I design and develop the Traffic Intersection Simulation System (TiSS), which can simulate any traffic signal methods, any intersections, regular or non-regular. It includes an optimal traffic signal method module, which can get the optimal signal method from the intersection and the method can be validated in TiSS. This paper describes the system structure design of TiSS, the traffic signal method module, the detector module and the system for the estimation of road cost function (travel time) parameters.